diff --git "a/checkpoint-1798/trainer_state.json" "b/checkpoint-1798/trainer_state.json"
new file mode 100644--- /dev/null
+++ "b/checkpoint-1798/trainer_state.json"
@@ -0,0 +1,12607 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 1.0,
+  "eval_steps": 500,
+  "global_step": 1798,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.0,
+      "grad_norm": 0.02996581234037876,
+      "learning_rate": 2e-05,
+      "loss": 1.2729,
+      "step": 1
+    },
+    {
+      "epoch": 0.0,
+      "grad_norm": 0.029612887650728226,
+      "learning_rate": 4e-05,
+      "loss": 1.208,
+      "step": 2
+    },
+    {
+      "epoch": 0.0,
+      "grad_norm": 0.027333809062838554,
+      "learning_rate": 6e-05,
+      "loss": 1.0525,
+      "step": 3
+    },
+    {
+      "epoch": 0.0,
+      "grad_norm": 0.02706489898264408,
+      "learning_rate": 8e-05,
+      "loss": 1.021,
+      "step": 4
+    },
+    {
+      "epoch": 0.0,
+      "grad_norm": 0.033770810812711716,
+      "learning_rate": 0.0001,
+      "loss": 1.223,
+      "step": 5
+    },
+    {
+      "epoch": 0.0,
+      "grad_norm": 0.03023155778646469,
+      "learning_rate": 0.00012,
+      "loss": 1.0973,
+      "step": 6
+    },
+    {
+      "epoch": 0.0,
+      "grad_norm": 0.034239206463098526,
+      "learning_rate": 0.00014,
+      "loss": 1.1578,
+      "step": 7
+    },
+    {
+      "epoch": 0.0,
+      "grad_norm": 0.037418968975543976,
+      "learning_rate": 0.00016,
+      "loss": 1.0797,
+      "step": 8
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 0.03651932254433632,
+      "learning_rate": 0.00018,
+      "loss": 1.0795,
+      "step": 9
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 0.03634285926818848,
+      "learning_rate": 0.0002,
+      "loss": 0.9999,
+      "step": 10
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 0.03892536088824272,
+      "learning_rate": 0.00019999996162493256,
+      "loss": 1.0103,
+      "step": 11
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 0.0294953566044569,
+      "learning_rate": 0.00019999984649975977,
+      "loss": 1.0729,
+      "step": 12
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 0.04408486932516098,
+      "learning_rate": 0.00019999965462456993,
+      "loss": 1.0439,
+      "step": 13
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 0.03400202468037605,
+      "learning_rate": 0.00019999938599951034,
+      "loss": 0.8412,
+      "step": 14
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 0.034877073019742966,
+      "learning_rate": 0.00019999904062478714,
+      "loss": 0.9438,
+      "step": 15
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 0.0340421162545681,
+      "learning_rate": 0.0001999986185006654,
+      "loss": 1.015,
+      "step": 16
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 0.03420884534716606,
+      "learning_rate": 0.00019999811962746915,
+      "loss": 0.9792,
+      "step": 17
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 0.03177829086780548,
+      "learning_rate": 0.00019999754400558124,
+      "loss": 0.8856,
+      "step": 18
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 0.03189516067504883,
+      "learning_rate": 0.00019999689163544346,
+      "loss": 0.8841,
+      "step": 19
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 0.025729015469551086,
+      "learning_rate": 0.00019999616251755651,
+      "loss": 0.9524,
+      "step": 20
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 0.02780766598880291,
+      "learning_rate": 0.00019999535665248002,
+      "loss": 0.8998,
+      "step": 21
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 0.023537974804639816,
+      "learning_rate": 0.00019999447404083244,
+      "loss": 0.9479,
+      "step": 22
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 0.023690495640039444,
+      "learning_rate": 0.0001999935146832912,
+      "loss": 0.9138,
+      "step": 23
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 0.023732604458928108,
+      "learning_rate": 0.00019999247858059257,
+      "loss": 1.036,
+      "step": 24
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 0.02902216836810112,
+      "learning_rate": 0.00019999136573353184,
+      "loss": 0.9619,
+      "step": 25
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 0.024132205173373222,
+      "learning_rate": 0.0001999901761429631,
+      "loss": 0.9338,
+      "step": 26
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 0.02214036136865616,
+      "learning_rate": 0.0001999889098097993,
+      "loss": 0.8229,
+      "step": 27
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 0.023357391357421875,
+      "learning_rate": 0.0001999875667350124,
+      "loss": 0.8983,
+      "step": 28
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 0.022214027121663094,
+      "learning_rate": 0.00019998614691963323,
+      "loss": 0.7894,
+      "step": 29
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 0.021391017362475395,
+      "learning_rate": 0.00019998465036475148,
+      "loss": 0.9419,
+      "step": 30
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 0.02627565711736679,
+      "learning_rate": 0.0001999830770715157,
+      "loss": 0.9285,
+      "step": 31
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 0.022590599954128265,
+      "learning_rate": 0.0001999814270411335,
+      "loss": 0.9401,
+      "step": 32
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 0.024948442354798317,
+      "learning_rate": 0.00019997970027487122,
+      "loss": 0.9069,
+      "step": 33
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 0.02103595621883869,
+      "learning_rate": 0.00019997789677405416,
+      "loss": 0.9892,
+      "step": 34
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 0.02005797065794468,
+      "learning_rate": 0.0001999760165400665,
+      "loss": 0.9577,
+      "step": 35
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 0.021683409810066223,
+      "learning_rate": 0.00019997405957435135,
+      "loss": 0.9099,
+      "step": 36
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 0.020573126152157784,
+      "learning_rate": 0.00019997202587841066,
+      "loss": 0.8264,
+      "step": 37
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 0.023178325966000557,
+      "learning_rate": 0.00019996991545380533,
+      "loss": 0.8658,
+      "step": 38
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 0.01991512067615986,
+      "learning_rate": 0.00019996772830215505,
+      "loss": 0.9239,
+      "step": 39
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 0.02087959088385105,
+      "learning_rate": 0.00019996546442513854,
+      "loss": 0.8532,
+      "step": 40
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 0.01916460692882538,
+      "learning_rate": 0.00019996312382449327,
+      "loss": 1.0405,
+      "step": 41
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 0.023638034239411354,
+      "learning_rate": 0.00019996070650201567,
+      "loss": 0.9984,
+      "step": 42
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 0.023328714072704315,
+      "learning_rate": 0.000199958212459561,
+      "loss": 0.8696,
+      "step": 43
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 0.021662859246134758,
+      "learning_rate": 0.00019995564169904354,
+      "loss": 0.8519,
+      "step": 44
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 0.024608176201581955,
+      "learning_rate": 0.0001999529942224363,
+      "loss": 0.8428,
+      "step": 45
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 0.020750664174556732,
+      "learning_rate": 0.00019995027003177118,
+      "loss": 0.8653,
+      "step": 46
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 0.021747557446360588,
+      "learning_rate": 0.00019994746912913906,
+      "loss": 0.8973,
+      "step": 47
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 0.02252582646906376,
+      "learning_rate": 0.00019994459151668957,
+      "loss": 0.8664,
+      "step": 48
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 0.0229500625282526,
+      "learning_rate": 0.00019994163719663137,
+      "loss": 0.8613,
+      "step": 49
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 0.021505488082766533,
+      "learning_rate": 0.00019993860617123184,
+      "loss": 0.812,
+      "step": 50
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 0.020829355344176292,
+      "learning_rate": 0.0001999354984428173,
+      "loss": 0.9139,
+      "step": 51
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 0.024170633405447006,
+      "learning_rate": 0.000199932314013773,
+      "loss": 0.937,
+      "step": 52
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 0.021097104996442795,
+      "learning_rate": 0.0001999290528865429,
+      "loss": 0.9093,
+      "step": 53
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 0.023191392421722412,
+      "learning_rate": 0.00019992571506363,
+      "loss": 0.9728,
+      "step": 54
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 0.023909488692879677,
+      "learning_rate": 0.000199922300547596,
+      "loss": 0.8279,
+      "step": 55
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 0.024314479902386665,
+      "learning_rate": 0.0001999188093410616,
+      "loss": 0.8118,
+      "step": 56
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 0.022049203515052795,
+      "learning_rate": 0.00019991524144670635,
+      "loss": 0.8637,
+      "step": 57
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 0.030246440321207047,
+      "learning_rate": 0.0001999115968672685,
+      "loss": 0.8515,
+      "step": 58
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 0.021488850936293602,
+      "learning_rate": 0.00019990787560554538,
+      "loss": 0.844,
+      "step": 59
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 0.022307060658931732,
+      "learning_rate": 0.00019990407766439297,
+      "loss": 0.7771,
+      "step": 60
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 0.027180947363376617,
+      "learning_rate": 0.00019990020304672628,
+      "loss": 0.9951,
+      "step": 61
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 0.02594638243317604,
+      "learning_rate": 0.000199896251755519,
+      "loss": 0.8643,
+      "step": 62
+    },
+    {
+      "epoch": 0.04,
+      "grad_norm": 0.025260182097554207,
+      "learning_rate": 0.00019989222379380384,
+      "loss": 0.9858,
+      "step": 63
+    },
+    {
+      "epoch": 0.04,
+      "grad_norm": 0.029186906293034554,
+      "learning_rate": 0.0001998881191646722,
+      "loss": 0.7514,
+      "step": 64
+    },
+    {
+      "epoch": 0.04,
+      "grad_norm": 0.02333957888185978,
+      "learning_rate": 0.00019988393787127441,
+      "loss": 0.7914,
+      "step": 65
+    },
+    {
+      "epoch": 0.04,
+      "grad_norm": 0.0235314778983593,
+      "learning_rate": 0.00019987967991681965,
+      "loss": 0.8198,
+      "step": 66
+    },
+    {
+      "epoch": 0.04,
+      "grad_norm": 0.023193303495645523,
+      "learning_rate": 0.00019987534530457583,
+      "loss": 0.8976,
+      "step": 67
+    },
+    {
+      "epoch": 0.04,
+      "grad_norm": 0.026709623634815216,
+      "learning_rate": 0.00019987093403786986,
+      "loss": 0.8833,
+      "step": 68
+    },
+    {
+      "epoch": 0.04,
+      "grad_norm": 0.024577545002102852,
+      "learning_rate": 0.00019986644612008728,
+      "loss": 0.8259,
+      "step": 69
+    },
+    {
+      "epoch": 0.04,
+      "grad_norm": 0.027963118627667427,
+      "learning_rate": 0.0001998618815546727,
+      "loss": 0.8767,
+      "step": 70
+    },
+    {
+      "epoch": 0.04,
+      "grad_norm": 0.02565196342766285,
+      "learning_rate": 0.00019985724034512936,
+      "loss": 0.7295,
+      "step": 71
+    },
+    {
+      "epoch": 0.04,
+      "grad_norm": 0.026185255497694016,
+      "learning_rate": 0.0001998525224950194,
+      "loss": 0.876,
+      "step": 72
+    },
+    {
+      "epoch": 0.04,
+      "grad_norm": 0.023345299065113068,
+      "learning_rate": 0.00019984772800796377,
+      "loss": 0.997,
+      "step": 73
+    },
+    {
+      "epoch": 0.04,
+      "grad_norm": 0.024012910202145576,
+      "learning_rate": 0.00019984285688764226,
+      "loss": 0.9337,
+      "step": 74
+    },
+    {
+      "epoch": 0.04,
+      "grad_norm": 0.02843514457345009,
+      "learning_rate": 0.00019983790913779347,
+      "loss": 0.8949,
+      "step": 75
+    },
+    {
+      "epoch": 0.04,
+      "grad_norm": 0.022462787106633186,
+      "learning_rate": 0.0001998328847622148,
+      "loss": 0.9018,
+      "step": 76
+    },
+    {
+      "epoch": 0.04,
+      "grad_norm": 0.024008071050047874,
+      "learning_rate": 0.00019982778376476245,
+      "loss": 0.8376,
+      "step": 77
+    },
+    {
+      "epoch": 0.04,
+      "grad_norm": 0.023886198177933693,
+      "learning_rate": 0.00019982260614935144,
+      "loss": 0.8118,
+      "step": 78
+    },
+    {
+      "epoch": 0.04,
+      "grad_norm": 0.025386439636349678,
+      "learning_rate": 0.00019981735191995563,
+      "loss": 0.9161,
+      "step": 79
+    },
+    {
+      "epoch": 0.04,
+      "grad_norm": 0.02613622136414051,
+      "learning_rate": 0.0001998120210806076,
+      "loss": 0.9008,
+      "step": 80
+    },
+    {
+      "epoch": 0.05,
+      "grad_norm": 0.02593972161412239,
+      "learning_rate": 0.00019980661363539883,
+      "loss": 0.9567,
+      "step": 81
+    },
+    {
+      "epoch": 0.05,
+      "grad_norm": 0.02825133316218853,
+      "learning_rate": 0.0001998011295884795,
+      "loss": 0.8386,
+      "step": 82
+    },
+    {
+      "epoch": 0.05,
+      "grad_norm": 0.025737276300787926,
+      "learning_rate": 0.00019979556894405862,
+      "loss": 0.7957,
+      "step": 83
+    },
+    {
+      "epoch": 0.05,
+      "grad_norm": 0.023754647001624107,
+      "learning_rate": 0.00019978993170640403,
+      "loss": 0.8167,
+      "step": 84
+    },
+    {
+      "epoch": 0.05,
+      "grad_norm": 0.02431335486471653,
+      "learning_rate": 0.00019978421787984228,
+      "loss": 0.7568,
+      "step": 85
+    },
+    {
+      "epoch": 0.05,
+      "grad_norm": 0.024608569219708443,
+      "learning_rate": 0.00019977842746875878,
+      "loss": 0.7992,
+      "step": 86
+    },
+    {
+      "epoch": 0.05,
+      "grad_norm": 0.024010393768548965,
+      "learning_rate": 0.00019977256047759765,
+      "loss": 0.9094,
+      "step": 87
+    },
+    {
+      "epoch": 0.05,
+      "grad_norm": 0.023806700482964516,
+      "learning_rate": 0.00019976661691086182,
+      "loss": 0.8677,
+      "step": 88
+    },
+    {
+      "epoch": 0.05,
+      "grad_norm": 0.02781900390982628,
+      "learning_rate": 0.00019976059677311297,
+      "loss": 0.7736,
+      "step": 89
+    },
+    {
+      "epoch": 0.05,
+      "grad_norm": 0.027234690263867378,
+      "learning_rate": 0.0001997545000689716,
+      "loss": 0.789,
+      "step": 90
+    },
+    {
+      "epoch": 0.05,
+      "grad_norm": 0.028865503147244453,
+      "learning_rate": 0.00019974832680311688,
+      "loss": 0.9041,
+      "step": 91
+    },
+    {
+      "epoch": 0.05,
+      "grad_norm": 0.05750074237585068,
+      "learning_rate": 0.00019974207698028685,
+      "loss": 0.8886,
+      "step": 92
+    },
+    {
+      "epoch": 0.05,
+      "grad_norm": 0.025579992681741714,
+      "learning_rate": 0.00019973575060527825,
+      "loss": 0.835,
+      "step": 93
+    },
+    {
+      "epoch": 0.05,
+      "grad_norm": 0.02479216828942299,
+      "learning_rate": 0.00019972934768294659,
+      "loss": 0.765,
+      "step": 94
+    },
+    {
+      "epoch": 0.05,
+      "grad_norm": 0.024221569299697876,
+      "learning_rate": 0.0001997228682182061,
+      "loss": 0.7935,
+      "step": 95
+    },
+    {
+      "epoch": 0.05,
+      "grad_norm": 0.025168055668473244,
+      "learning_rate": 0.00019971631221602976,
+      "loss": 0.7801,
+      "step": 96
+    },
+    {
+      "epoch": 0.05,
+      "grad_norm": 0.027467774227261543,
+      "learning_rate": 0.00019970967968144937,
+      "loss": 0.8287,
+      "step": 97
+    },
+    {
+      "epoch": 0.05,
+      "grad_norm": 0.023378336802124977,
+      "learning_rate": 0.00019970297061955533,
+      "loss": 0.7486,
+      "step": 98
+    },
+    {
+      "epoch": 0.06,
+      "grad_norm": 0.025852292776107788,
+      "learning_rate": 0.00019969618503549693,
+      "loss": 0.8939,
+      "step": 99
+    },
+    {
+      "epoch": 0.06,
+      "grad_norm": 0.02506246417760849,
+      "learning_rate": 0.00019968932293448207,
+      "loss": 0.8112,
+      "step": 100
+    },
+    {
+      "epoch": 0.06,
+      "grad_norm": 0.026856260374188423,
+      "learning_rate": 0.00019968238432177744,
+      "loss": 0.8782,
+      "step": 101
+    },
+    {
+      "epoch": 0.06,
+      "grad_norm": 0.028643622994422913,
+      "learning_rate": 0.0001996753692027084,
+      "loss": 0.7884,
+      "step": 102
+    },
+    {
+      "epoch": 0.06,
+      "grad_norm": 0.029276320710778236,
+      "learning_rate": 0.00019966827758265913,
+      "loss": 0.8926,
+      "step": 103
+    },
+    {
+      "epoch": 0.06,
+      "grad_norm": 0.027838734909892082,
+      "learning_rate": 0.00019966110946707244,
+      "loss": 0.87,
+      "step": 104
+    },
+    {
+      "epoch": 0.06,
+      "grad_norm": 0.026706190779805183,
+      "learning_rate": 0.0001996538648614498,
+      "loss": 0.9119,
+      "step": 105
+    },
+    {
+      "epoch": 0.06,
+      "grad_norm": 0.024687886238098145,
+      "learning_rate": 0.00019964654377135153,
+      "loss": 0.7647,
+      "step": 106
+    },
+    {
+      "epoch": 0.06,
+      "grad_norm": 0.02482466958463192,
+      "learning_rate": 0.00019963914620239656,
+      "loss": 0.894,
+      "step": 107
+    },
+    {
+      "epoch": 0.06,
+      "grad_norm": 0.024776702746748924,
+      "learning_rate": 0.0001996316721602625,
+      "loss": 0.7515,
+      "step": 108
+    },
+    {
+      "epoch": 0.06,
+      "grad_norm": 0.02663383074104786,
+      "learning_rate": 0.00019962412165068573,
+      "loss": 0.8499,
+      "step": 109
+    },
+    {
+      "epoch": 0.06,
+      "grad_norm": 0.027590934187173843,
+      "learning_rate": 0.00019961649467946125,
+      "loss": 0.7446,
+      "step": 110
+    },
+    {
+      "epoch": 0.06,
+      "grad_norm": 0.025346562266349792,
+      "learning_rate": 0.0001996087912524428,
+      "loss": 0.8308,
+      "step": 111
+    },
+    {
+      "epoch": 0.06,
+      "grad_norm": 0.02552192285656929,
+      "learning_rate": 0.0001996010113755427,
+      "loss": 0.8552,
+      "step": 112
+    },
+    {
+      "epoch": 0.06,
+      "grad_norm": 0.024720516055822372,
+      "learning_rate": 0.0001995931550547321,
+      "loss": 0.8648,
+      "step": 113
+    },
+    {
+      "epoch": 0.06,
+      "grad_norm": 0.02904409170150757,
+      "learning_rate": 0.0001995852222960407,
+      "loss": 0.8826,
+      "step": 114
+    },
+    {
+      "epoch": 0.06,
+      "grad_norm": 0.027031444013118744,
+      "learning_rate": 0.00019957721310555693,
+      "loss": 0.8748,
+      "step": 115
+    },
+    {
+      "epoch": 0.06,
+      "grad_norm": 0.025683356449007988,
+      "learning_rate": 0.0001995691274894278,
+      "loss": 0.8441,
+      "step": 116
+    },
+    {
+      "epoch": 0.07,
+      "grad_norm": 0.03195862099528313,
+      "learning_rate": 0.00019956096545385905,
+      "loss": 0.7829,
+      "step": 117
+    },
+    {
+      "epoch": 0.07,
+      "grad_norm": 0.025551646947860718,
+      "learning_rate": 0.00019955272700511507,
+      "loss": 0.7934,
+      "step": 118
+    },
+    {
+      "epoch": 0.07,
+      "grad_norm": 0.028811989352107048,
+      "learning_rate": 0.0001995444121495189,
+      "loss": 0.8001,
+      "step": 119
+    },
+    {
+      "epoch": 0.07,
+      "grad_norm": 0.030996061861515045,
+      "learning_rate": 0.00019953602089345217,
+      "loss": 0.7128,
+      "step": 120
+    },
+    {
+      "epoch": 0.07,
+      "grad_norm": 0.026600060984492302,
+      "learning_rate": 0.00019952755324335514,
+      "loss": 0.7414,
+      "step": 121
+    },
+    {
+      "epoch": 0.07,
+      "grad_norm": 0.025535378605127335,
+      "learning_rate": 0.00019951900920572684,
+      "loss": 0.8381,
+      "step": 122
+    },
+    {
+      "epoch": 0.07,
+      "grad_norm": 0.035039834678173065,
+      "learning_rate": 0.00019951038878712475,
+      "loss": 0.9607,
+      "step": 123
+    },
+    {
+      "epoch": 0.07,
+      "grad_norm": 0.03230242058634758,
+      "learning_rate": 0.00019950169199416513,
+      "loss": 0.876,
+      "step": 124
+    },
+    {
+      "epoch": 0.07,
+      "grad_norm": 0.031107056885957718,
+      "learning_rate": 0.0001994929188335227,
+      "loss": 0.8444,
+      "step": 125
+    },
+    {
+      "epoch": 0.07,
+      "grad_norm": 0.031076829880475998,
+      "learning_rate": 0.0001994840693119309,
+      "loss": 0.8577,
+      "step": 126
+    },
+    {
+      "epoch": 0.07,
+      "grad_norm": 0.028398986905813217,
+      "learning_rate": 0.0001994751434361818,
+      "loss": 0.7828,
+      "step": 127
+    },
+    {
+      "epoch": 0.07,
+      "grad_norm": 0.027821028605103493,
+      "learning_rate": 0.00019946614121312594,
+      "loss": 0.8243,
+      "step": 128
+    },
+    {
+      "epoch": 0.07,
+      "grad_norm": 0.030113285407423973,
+      "learning_rate": 0.0001994570626496726,
+      "loss": 0.7923,
+      "step": 129
+    },
+    {
+      "epoch": 0.07,
+      "grad_norm": 0.028168270364403725,
+      "learning_rate": 0.00019944790775278955,
+      "loss": 0.722,
+      "step": 130
+    },
+    {
+      "epoch": 0.07,
+      "grad_norm": 0.030299954116344452,
+      "learning_rate": 0.0001994386765295032,
+      "loss": 0.8462,
+      "step": 131
+    },
+    {
+      "epoch": 0.07,
+      "grad_norm": 0.029045602306723595,
+      "learning_rate": 0.00019942936898689854,
+      "loss": 0.7916,
+      "step": 132
+    },
+    {
+      "epoch": 0.07,
+      "grad_norm": 0.02803182788193226,
+      "learning_rate": 0.0001994199851321191,
+      "loss": 0.7949,
+      "step": 133
+    },
+    {
+      "epoch": 0.07,
+      "grad_norm": 0.029968315735459328,
+      "learning_rate": 0.00019941052497236703,
+      "loss": 0.9672,
+      "step": 134
+    },
+    {
+      "epoch": 0.08,
+      "grad_norm": 0.027893947437405586,
+      "learning_rate": 0.00019940098851490298,
+      "loss": 0.8284,
+      "step": 135
+    },
+    {
+      "epoch": 0.08,
+      "grad_norm": 0.030356327071785927,
+      "learning_rate": 0.0001993913757670462,
+      "loss": 0.8001,
+      "step": 136
+    },
+    {
+      "epoch": 0.08,
+      "grad_norm": 0.02701139822602272,
+      "learning_rate": 0.00019938168673617454,
+      "loss": 0.8137,
+      "step": 137
+    },
+    {
+      "epoch": 0.08,
+      "grad_norm": 0.032546404749155045,
+      "learning_rate": 0.00019937192142972427,
+      "loss": 0.7451,
+      "step": 138
+    },
+    {
+      "epoch": 0.08,
+      "grad_norm": 0.02752624824643135,
+      "learning_rate": 0.0001993620798551903,
+      "loss": 0.9893,
+      "step": 139
+    },
+    {
+      "epoch": 0.08,
+      "grad_norm": 0.02512490376830101,
+      "learning_rate": 0.0001993521620201261,
+      "loss": 0.9108,
+      "step": 140
+    },
+    {
+      "epoch": 0.08,
+      "grad_norm": 0.029267175123095512,
+      "learning_rate": 0.00019934216793214354,
+      "loss": 0.8274,
+      "step": 141
+    },
+    {
+      "epoch": 0.08,
+      "grad_norm": 0.02983001247048378,
+      "learning_rate": 0.00019933209759891317,
+      "loss": 0.8394,
+      "step": 142
+    },
+    {
+      "epoch": 0.08,
+      "grad_norm": 0.031017007306218147,
+      "learning_rate": 0.0001993219510281639,
+      "loss": 0.8455,
+      "step": 143
+    },
+    {
+      "epoch": 0.08,
+      "grad_norm": 0.02880048379302025,
+      "learning_rate": 0.00019931172822768335,
+      "loss": 0.8257,
+      "step": 144
+    },
+    {
+      "epoch": 0.08,
+      "grad_norm": 0.03203625977039337,
+      "learning_rate": 0.00019930142920531743,
+      "loss": 0.8693,
+      "step": 145
+    },
+    {
+      "epoch": 0.08,
+      "grad_norm": 0.027382776141166687,
+      "learning_rate": 0.0001992910539689707,
+      "loss": 0.8478,
+      "step": 146
+    },
+    {
+      "epoch": 0.08,
+      "grad_norm": 0.03054049052298069,
+      "learning_rate": 0.00019928060252660617,
+      "loss": 0.9593,
+      "step": 147
+    },
+    {
+      "epoch": 0.08,
+      "grad_norm": 0.026648158207535744,
+      "learning_rate": 0.00019927007488624535,
+      "loss": 0.8915,
+      "step": 148
+    },
+    {
+      "epoch": 0.08,
+      "grad_norm": 0.03315131738781929,
+      "learning_rate": 0.00019925947105596816,
+      "loss": 0.9008,
+      "step": 149
+    },
+    {
+      "epoch": 0.08,
+      "grad_norm": 0.027361949905753136,
+      "learning_rate": 0.00019924879104391309,
+      "loss": 0.8705,
+      "step": 150
+    },
+    {
+      "epoch": 0.08,
+      "grad_norm": 0.029892416670918465,
+      "learning_rate": 0.0001992380348582771,
+      "loss": 0.9083,
+      "step": 151
+    },
+    {
+      "epoch": 0.08,
+      "grad_norm": 0.027648137882351875,
+      "learning_rate": 0.00019922720250731553,
+      "loss": 0.7381,
+      "step": 152
+    },
+    {
+      "epoch": 0.09,
+      "grad_norm": 0.02656717598438263,
+      "learning_rate": 0.00019921629399934223,
+      "loss": 0.9495,
+      "step": 153
+    },
+    {
+      "epoch": 0.09,
+      "grad_norm": 0.027288291603326797,
+      "learning_rate": 0.0001992053093427295,
+      "loss": 0.8505,
+      "step": 154
+    },
+    {
+      "epoch": 0.09,
+      "grad_norm": 0.032024212181568146,
+      "learning_rate": 0.0001991942485459081,
+      "loss": 0.8288,
+      "step": 155
+    },
+    {
+      "epoch": 0.09,
+      "grad_norm": 0.027418840676546097,
+      "learning_rate": 0.00019918311161736717,
+      "loss": 0.8137,
+      "step": 156
+    },
+    {
+      "epoch": 0.09,
+      "grad_norm": 0.03174857795238495,
+      "learning_rate": 0.0001991718985656543,
+      "loss": 0.8754,
+      "step": 157
+    },
+    {
+      "epoch": 0.09,
+      "grad_norm": 0.02887050248682499,
+      "learning_rate": 0.0001991606093993756,
+      "loss": 0.8497,
+      "step": 158
+    },
+    {
+      "epoch": 0.09,
+      "grad_norm": 0.03064458630979061,
+      "learning_rate": 0.00019914924412719546,
+      "loss": 0.9331,
+      "step": 159
+    },
+    {
+      "epoch": 0.09,
+      "grad_norm": 0.02989426627755165,
+      "learning_rate": 0.00019913780275783674,
+      "loss": 0.9075,
+      "step": 160
+    },
+    {
+      "epoch": 0.09,
+      "grad_norm": 0.03449084982275963,
+      "learning_rate": 0.00019912628530008075,
+      "loss": 0.8531,
+      "step": 161
+    },
+    {
+      "epoch": 0.09,
+      "grad_norm": 0.03213748335838318,
+      "learning_rate": 0.00019911469176276712,
+      "loss": 0.7411,
+      "step": 162
+    },
+    {
+      "epoch": 0.09,
+      "grad_norm": 0.031524017453193665,
+      "learning_rate": 0.0001991030221547939,
+      "loss": 0.8003,
+      "step": 163
+    },
+    {
+      "epoch": 0.09,
+      "grad_norm": 0.028880394995212555,
+      "learning_rate": 0.00019909127648511755,
+      "loss": 0.86,
+      "step": 164
+    },
+    {
+      "epoch": 0.09,
+      "grad_norm": 0.0295130405575037,
+      "learning_rate": 0.0001990794547627529,
+      "loss": 0.802,
+      "step": 165
+    },
+    {
+      "epoch": 0.09,
+      "grad_norm": 0.032397400587797165,
+      "learning_rate": 0.00019906755699677312,
+      "loss": 0.8574,
+      "step": 166
+    },
+    {
+      "epoch": 0.09,
+      "grad_norm": 0.030780531466007233,
+      "learning_rate": 0.00019905558319630978,
+      "loss": 0.8748,
+      "step": 167
+    },
+    {
+      "epoch": 0.09,
+      "grad_norm": 0.026546768844127655,
+      "learning_rate": 0.0001990435333705527,
+      "loss": 0.8099,
+      "step": 168
+    },
+    {
+      "epoch": 0.09,
+      "grad_norm": 0.03083784319460392,
+      "learning_rate": 0.00019903140752875026,
+      "loss": 0.7867,
+      "step": 169
+    },
+    {
+      "epoch": 0.09,
+      "grad_norm": 0.02955833077430725,
+      "learning_rate": 0.000199019205680209,
+      "loss": 0.8806,
+      "step": 170
+    },
+    {
+      "epoch": 0.1,
+      "grad_norm": 0.029685625806450844,
+      "learning_rate": 0.00019900692783429385,
+      "loss": 0.9009,
+      "step": 171
+    },
+    {
+      "epoch": 0.1,
+      "grad_norm": 0.03194108605384827,
+      "learning_rate": 0.00019899457400042808,
+      "loss": 0.7654,
+      "step": 172
+    },
+    {
+      "epoch": 0.1,
+      "grad_norm": 0.032116640359163284,
+      "learning_rate": 0.0001989821441880933,
+      "loss": 1.0055,
+      "step": 173
+    },
+    {
+      "epoch": 0.1,
+      "grad_norm": 0.03061555325984955,
+      "learning_rate": 0.00019896963840682937,
+      "loss": 0.782,
+      "step": 174
+    },
+    {
+      "epoch": 0.1,
+      "grad_norm": 0.030177190899848938,
+      "learning_rate": 0.0001989570566662345,
+      "loss": 0.7403,
+      "step": 175
+    },
+    {
+      "epoch": 0.1,
+      "grad_norm": 0.031266387552022934,
+      "learning_rate": 0.00019894439897596522,
+      "loss": 0.8553,
+      "step": 176
+    },
+    {
+      "epoch": 0.1,
+      "grad_norm": 0.032219842076301575,
+      "learning_rate": 0.0001989316653457363,
+      "loss": 0.7479,
+      "step": 177
+    },
+    {
+      "epoch": 0.1,
+      "grad_norm": 0.029450630769133568,
+      "learning_rate": 0.0001989188557853208,
+      "loss": 0.7373,
+      "step": 178
+    },
+    {
+      "epoch": 0.1,
+      "grad_norm": 0.03027605637907982,
+      "learning_rate": 0.00019890597030455015,
+      "loss": 0.7844,
+      "step": 179
+    },
+    {
+      "epoch": 0.1,
+      "grad_norm": 0.03511941432952881,
+      "learning_rate": 0.00019889300891331392,
+      "loss": 0.8679,
+      "step": 180
+    },
+    {
+      "epoch": 0.1,
+      "grad_norm": 0.031290799379348755,
+      "learning_rate": 0.00019887997162155996,
+      "loss": 0.7283,
+      "step": 181
+    },
+    {
+      "epoch": 0.1,
+      "grad_norm": 0.02689247764647007,
+      "learning_rate": 0.0001988668584392945,
+      "loss": 0.7628,
+      "step": 182
+    },
+    {
+      "epoch": 0.1,
+      "grad_norm": 0.030273688957095146,
+      "learning_rate": 0.0001988536693765818,
+      "loss": 0.7985,
+      "step": 183
+    },
+    {
+      "epoch": 0.1,
+      "grad_norm": 0.030390895903110504,
+      "learning_rate": 0.00019884040444354461,
+      "loss": 0.8086,
+      "step": 184
+    },
+    {
+      "epoch": 0.1,
+      "grad_norm": 0.029167182743549347,
+      "learning_rate": 0.00019882706365036373,
+      "loss": 0.7081,
+      "step": 185
+    },
+    {
+      "epoch": 0.1,
+      "grad_norm": 0.03029320389032364,
+      "learning_rate": 0.00019881364700727823,
+      "loss": 0.6738,
+      "step": 186
+    },
+    {
+      "epoch": 0.1,
+      "grad_norm": 0.028749890625476837,
+      "learning_rate": 0.00019880015452458544,
+      "loss": 0.812,
+      "step": 187
+    },
+    {
+      "epoch": 0.1,
+      "grad_norm": 0.030206618830561638,
+      "learning_rate": 0.0001987865862126408,
+      "loss": 0.8358,
+      "step": 188
+    },
+    {
+      "epoch": 0.11,
+      "grad_norm": 0.030770782381296158,
+      "learning_rate": 0.00019877294208185803,
+      "loss": 0.7326,
+      "step": 189
+    },
+    {
+      "epoch": 0.11,
+      "grad_norm": 0.03188487887382507,
+      "learning_rate": 0.00019875922214270903,
+      "loss": 0.817,
+      "step": 190
+    },
+    {
+      "epoch": 0.11,
+      "grad_norm": 0.031357571482658386,
+      "learning_rate": 0.00019874542640572384,
+      "loss": 0.7981,
+      "step": 191
+    },
+    {
+      "epoch": 0.11,
+      "grad_norm": 0.06510663032531738,
+      "learning_rate": 0.00019873155488149078,
+      "loss": 0.8416,
+      "step": 192
+    },
+    {
+      "epoch": 0.11,
+      "grad_norm": 0.030021656304597855,
+      "learning_rate": 0.0001987176075806562,
+      "loss": 0.7412,
+      "step": 193
+    },
+    {
+      "epoch": 0.11,
+      "grad_norm": 0.033732689917087555,
+      "learning_rate": 0.00019870358451392467,
+      "loss": 1.0041,
+      "step": 194
+    },
+    {
+      "epoch": 0.11,
+      "grad_norm": 0.02999632991850376,
+      "learning_rate": 0.00019868948569205897,
+      "loss": 0.7975,
+      "step": 195
+    },
+    {
+      "epoch": 0.11,
+      "grad_norm": 0.02894209697842598,
+      "learning_rate": 0.0001986753111258799,
+      "loss": 0.7465,
+      "step": 196
+    },
+    {
+      "epoch": 0.11,
+      "grad_norm": 0.033549994230270386,
+      "learning_rate": 0.0001986610608262665,
+      "loss": 0.8114,
+      "step": 197
+    },
+    {
+      "epoch": 0.11,
+      "grad_norm": 0.03032383695244789,
+      "learning_rate": 0.00019864673480415589,
+      "loss": 0.7155,
+      "step": 198
+    },
+    {
+      "epoch": 0.11,
+      "grad_norm": 0.03509935364127159,
+      "learning_rate": 0.00019863233307054325,
+      "loss": 0.8527,
+      "step": 199
+    },
+    {
+      "epoch": 0.11,
+      "grad_norm": 0.03238251060247421,
+      "learning_rate": 0.00019861785563648202,
+      "loss": 0.7574,
+      "step": 200
+    },
+    {
+      "epoch": 0.11,
+      "grad_norm": 0.03257328271865845,
+      "learning_rate": 0.00019860330251308362,
+      "loss": 0.9374,
+      "step": 201
+    },
+    {
+      "epoch": 0.11,
+      "grad_norm": 0.03191230073571205,
+      "learning_rate": 0.00019858867371151754,
+      "loss": 0.8127,
+      "step": 202
+    },
+    {
+      "epoch": 0.11,
+      "grad_norm": 0.030960606411099434,
+      "learning_rate": 0.0001985739692430115,
+      "loss": 0.7888,
+      "step": 203
+    },
+    {
+      "epoch": 0.11,
+      "grad_norm": 0.03744737431406975,
+      "learning_rate": 0.0001985591891188511,
+      "loss": 0.8152,
+      "step": 204
+    },
+    {
+      "epoch": 0.11,
+      "grad_norm": 0.03747240826487541,
+      "learning_rate": 0.0001985443333503802,
+      "loss": 0.7457,
+      "step": 205
+    },
+    {
+      "epoch": 0.11,
+      "grad_norm": 0.030524997040629387,
+      "learning_rate": 0.00019852940194900053,
+      "loss": 0.8093,
+      "step": 206
+    },
+    {
+      "epoch": 0.12,
+      "grad_norm": 0.032599907368421555,
+      "learning_rate": 0.00019851439492617204,
+      "loss": 0.7419,
+      "step": 207
+    },
+    {
+      "epoch": 0.12,
+      "grad_norm": 0.0342191606760025,
+      "learning_rate": 0.00019849931229341258,
+      "loss": 0.7183,
+      "step": 208
+    },
+    {
+      "epoch": 0.12,
+      "grad_norm": 0.03359980136156082,
+      "learning_rate": 0.00019848415406229812,
+      "loss": 0.8392,
+      "step": 209
+    },
+    {
+      "epoch": 0.12,
+      "grad_norm": 0.03139995038509369,
+      "learning_rate": 0.00019846892024446265,
+      "loss": 0.7657,
+      "step": 210
+    },
+    {
+      "epoch": 0.12,
+      "grad_norm": 0.029806002974510193,
+      "learning_rate": 0.00019845361085159806,
+      "loss": 0.6918,
+      "step": 211
+    },
+    {
+      "epoch": 0.12,
+      "grad_norm": 0.03195658698678017,
+      "learning_rate": 0.00019843822589545441,
+      "loss": 0.8361,
+      "step": 212
+    },
+    {
+      "epoch": 0.12,
+      "grad_norm": 0.03134671226143837,
+      "learning_rate": 0.00019842276538783966,
+      "loss": 0.9247,
+      "step": 213
+    },
+    {
+      "epoch": 0.12,
+      "grad_norm": 0.03182549774646759,
+      "learning_rate": 0.00019840722934061974,
+      "loss": 0.8478,
+      "step": 214
+    },
+    {
+      "epoch": 0.12,
+      "grad_norm": 0.035036999732255936,
+      "learning_rate": 0.00019839161776571864,
+      "loss": 0.7699,
+      "step": 215
+    },
+    {
+      "epoch": 0.12,
+      "grad_norm": 0.030924562364816666,
+      "learning_rate": 0.00019837593067511823,
+      "loss": 0.8543,
+      "step": 216
+    },
+    {
+      "epoch": 0.12,
+      "grad_norm": 0.03198299929499626,
+      "learning_rate": 0.00019836016808085836,
+      "loss": 0.761,
+      "step": 217
+    },
+    {
+      "epoch": 0.12,
+      "grad_norm": 0.03266490250825882,
+      "learning_rate": 0.00019834432999503684,
+      "loss": 0.7841,
+      "step": 218
+    },
+    {
+      "epoch": 0.12,
+      "grad_norm": 0.03606751561164856,
+      "learning_rate": 0.00019832841642980945,
+      "loss": 0.8398,
+      "step": 219
+    },
+    {
+      "epoch": 0.12,
+      "grad_norm": 0.03128010407090187,
+      "learning_rate": 0.00019831242739738985,
+      "loss": 0.7795,
+      "step": 220
+    },
+    {
+      "epoch": 0.12,
+      "grad_norm": 0.031938567757606506,
+      "learning_rate": 0.00019829636291004968,
+      "loss": 0.7383,
+      "step": 221
+    },
+    {
+      "epoch": 0.12,
+      "grad_norm": 0.03146064281463623,
+      "learning_rate": 0.00019828022298011842,
+      "loss": 0.8463,
+      "step": 222
+    },
+    {
+      "epoch": 0.12,
+      "grad_norm": 0.03626205027103424,
+      "learning_rate": 0.00019826400761998353,
+      "loss": 0.9204,
+      "step": 223
+    },
+    {
+      "epoch": 0.12,
+      "grad_norm": 0.03110690787434578,
+      "learning_rate": 0.00019824771684209025,
+      "loss": 0.7676,
+      "step": 224
+    },
+    {
+      "epoch": 0.13,
+      "grad_norm": 0.0522528775036335,
+      "learning_rate": 0.00019823135065894185,
+      "loss": 0.8221,
+      "step": 225
+    },
+    {
+      "epoch": 0.13,
+      "grad_norm": 0.031512025743722916,
+      "learning_rate": 0.00019821490908309932,
+      "loss": 0.745,
+      "step": 226
+    },
+    {
+      "epoch": 0.13,
+      "grad_norm": 0.03326869383454323,
+      "learning_rate": 0.00019819839212718168,
+      "loss": 0.8885,
+      "step": 227
+    },
+    {
+      "epoch": 0.13,
+      "grad_norm": 0.03730284795165062,
+      "learning_rate": 0.00019818179980386563,
+      "loss": 0.7942,
+      "step": 228
+    },
+    {
+      "epoch": 0.13,
+      "grad_norm": 0.03505445271730423,
+      "learning_rate": 0.00019816513212588585,
+      "loss": 0.7352,
+      "step": 229
+    },
+    {
+      "epoch": 0.13,
+      "grad_norm": 0.03741452097892761,
+      "learning_rate": 0.00019814838910603481,
+      "loss": 0.9211,
+      "step": 230
+    },
+    {
+      "epoch": 0.13,
+      "grad_norm": 0.03283902630209923,
+      "learning_rate": 0.00019813157075716277,
+      "loss": 0.9182,
+      "step": 231
+    },
+    {
+      "epoch": 0.13,
+      "grad_norm": 0.0364568717777729,
+      "learning_rate": 0.0001981146770921779,
+      "loss": 0.9211,
+      "step": 232
+    },
+    {
+      "epoch": 0.13,
+      "grad_norm": 0.03377554938197136,
+      "learning_rate": 0.000198097708124046,
+      "loss": 0.8209,
+      "step": 233
+    },
+    {
+      "epoch": 0.13,
+      "grad_norm": 0.030919229611754417,
+      "learning_rate": 0.00019808066386579085,
+      "loss": 0.7564,
+      "step": 234
+    },
+    {
+      "epoch": 0.13,
+      "grad_norm": 0.032216716557741165,
+      "learning_rate": 0.00019806354433049393,
+      "loss": 0.8913,
+      "step": 235
+    },
+    {
+      "epoch": 0.13,
+      "grad_norm": 0.038493942469358444,
+      "learning_rate": 0.00019804634953129454,
+      "loss": 0.701,
+      "step": 236
+    },
+    {
+      "epoch": 0.13,
+      "grad_norm": 0.047857046127319336,
+      "learning_rate": 0.00019802907948138963,
+      "loss": 1.0259,
+      "step": 237
+    },
+    {
+      "epoch": 0.13,
+      "grad_norm": 0.0336415059864521,
+      "learning_rate": 0.00019801173419403405,
+      "loss": 0.9172,
+      "step": 238
+    },
+    {
+      "epoch": 0.13,
+      "grad_norm": 0.030060969293117523,
+      "learning_rate": 0.00019799431368254034,
+      "loss": 0.8397,
+      "step": 239
+    },
+    {
+      "epoch": 0.13,
+      "grad_norm": 0.043135106563568115,
+      "learning_rate": 0.0001979768179602787,
+      "loss": 0.9176,
+      "step": 240
+    },
+    {
+      "epoch": 0.13,
+      "grad_norm": 0.03419271484017372,
+      "learning_rate": 0.00019795924704067721,
+      "loss": 0.9116,
+      "step": 241
+    },
+    {
+      "epoch": 0.13,
+      "grad_norm": 0.03154682740569115,
+      "learning_rate": 0.0001979416009372215,
+      "loss": 0.7786,
+      "step": 242
+    },
+    {
+      "epoch": 0.14,
+      "grad_norm": 0.03246486559510231,
+      "learning_rate": 0.000197923879663455,
+      "loss": 0.7744,
+      "step": 243
+    },
+    {
+      "epoch": 0.14,
+      "grad_norm": 0.03503431752324104,
+      "learning_rate": 0.0001979060832329788,
+      "loss": 0.7365,
+      "step": 244
+    },
+    {
+      "epoch": 0.14,
+      "grad_norm": 0.03330378606915474,
+      "learning_rate": 0.00019788821165945172,
+      "loss": 0.6759,
+      "step": 245
+    },
+    {
+      "epoch": 0.14,
+      "grad_norm": 0.030394921079277992,
+      "learning_rate": 0.00019787026495659022,
+      "loss": 0.7804,
+      "step": 246
+    },
+    {
+      "epoch": 0.14,
+      "grad_norm": 0.032463330775499344,
+      "learning_rate": 0.00019785224313816836,
+      "loss": 0.8431,
+      "step": 247
+    },
+    {
+      "epoch": 0.14,
+      "grad_norm": 0.032576996833086014,
+      "learning_rate": 0.00019783414621801798,
+      "loss": 0.787,
+      "step": 248
+    },
+    {
+      "epoch": 0.14,
+      "grad_norm": 0.03479291498661041,
+      "learning_rate": 0.00019781597421002843,
+      "loss": 0.8343,
+      "step": 249
+    },
+    {
+      "epoch": 0.14,
+      "grad_norm": 0.03418460860848427,
+      "learning_rate": 0.00019779772712814677,
+      "loss": 0.8228,
+      "step": 250
+    },
+    {
+      "epoch": 0.14,
+      "grad_norm": 0.03314866125583649,
+      "learning_rate": 0.0001977794049863777,
+      "loss": 0.8703,
+      "step": 251
+    },
+    {
+      "epoch": 0.14,
+      "grad_norm": 0.031179893761873245,
+      "learning_rate": 0.00019776100779878345,
+      "loss": 0.6981,
+      "step": 252
+    },
+    {
+      "epoch": 0.14,
+      "grad_norm": 0.03774740546941757,
+      "learning_rate": 0.00019774253557948386,
+      "loss": 0.5688,
+      "step": 253
+    },
+    {
+      "epoch": 0.14,
+      "grad_norm": 0.033188916742801666,
+      "learning_rate": 0.00019772398834265643,
+      "loss": 0.7994,
+      "step": 254
+    },
+    {
+      "epoch": 0.14,
+      "grad_norm": 0.03498241677880287,
+      "learning_rate": 0.00019770536610253618,
+      "loss": 0.7059,
+      "step": 255
+    },
+    {
+      "epoch": 0.14,
+      "grad_norm": 0.035760391503572464,
+      "learning_rate": 0.00019768666887341567,
+      "loss": 0.8468,
+      "step": 256
+    },
+    {
+      "epoch": 0.14,
+      "grad_norm": 0.0330347940325737,
+      "learning_rate": 0.0001976678966696451,
+      "loss": 0.7864,
+      "step": 257
+    },
+    {
+      "epoch": 0.14,
+      "grad_norm": 0.036673109978437424,
+      "learning_rate": 0.00019764904950563214,
+      "loss": 0.867,
+      "step": 258
+    },
+    {
+      "epoch": 0.14,
+      "grad_norm": 0.03514648601412773,
+      "learning_rate": 0.00019763012739584205,
+      "loss": 0.9021,
+      "step": 259
+    },
+    {
+      "epoch": 0.14,
+      "grad_norm": 0.03706090897321701,
+      "learning_rate": 0.00019761113035479747,
+      "loss": 0.7132,
+      "step": 260
+    },
+    {
+      "epoch": 0.15,
+      "grad_norm": 0.033022210001945496,
+      "learning_rate": 0.00019759205839707877,
+      "loss": 0.8326,
+      "step": 261
+    },
+    {
+      "epoch": 0.15,
+      "grad_norm": 0.03423427790403366,
+      "learning_rate": 0.00019757291153732362,
+      "loss": 0.9028,
+      "step": 262
+    },
+    {
+      "epoch": 0.15,
+      "grad_norm": 0.03209824860095978,
+      "learning_rate": 0.00019755368979022732,
+      "loss": 0.7991,
+      "step": 263
+    },
+    {
+      "epoch": 0.15,
+      "grad_norm": 0.03312317654490471,
+      "learning_rate": 0.00019753439317054256,
+      "loss": 0.7846,
+      "step": 264
+    },
+    {
+      "epoch": 0.15,
+      "grad_norm": 0.03310060128569603,
+      "learning_rate": 0.00019751502169307954,
+      "loss": 0.7369,
+      "step": 265
+    },
+    {
+      "epoch": 0.15,
+      "grad_norm": 0.03352689370512962,
+      "learning_rate": 0.00019749557537270584,
+      "loss": 0.8118,
+      "step": 266
+    },
+    {
+      "epoch": 0.15,
+      "grad_norm": 0.037001077085733414,
+      "learning_rate": 0.00019747605422434662,
+      "loss": 0.7978,
+      "step": 267
+    },
+    {
+      "epoch": 0.15,
+      "grad_norm": 0.03153040632605553,
+      "learning_rate": 0.00019745645826298432,
+      "loss": 0.6908,
+      "step": 268
+    },
+    {
+      "epoch": 0.15,
+      "grad_norm": 0.029728813096880913,
+      "learning_rate": 0.00019743678750365888,
+      "loss": 0.7491,
+      "step": 269
+    },
+    {
+      "epoch": 0.15,
+      "grad_norm": 0.04103512689471245,
+      "learning_rate": 0.00019741704196146766,
+      "loss": 0.8505,
+      "step": 270
+    },
+    {
+      "epoch": 0.15,
+      "grad_norm": 0.036623742431402206,
+      "learning_rate": 0.00019739722165156538,
+      "loss": 0.8811,
+      "step": 271
+    },
+    {
+      "epoch": 0.15,
+      "grad_norm": 0.031018711626529694,
+      "learning_rate": 0.00019737732658916412,
+      "loss": 0.7914,
+      "step": 272
+    },
+    {
+      "epoch": 0.15,
+      "grad_norm": 0.0346776582300663,
+      "learning_rate": 0.00019735735678953344,
+      "loss": 0.8786,
+      "step": 273
+    },
+    {
+      "epoch": 0.15,
+      "grad_norm": 0.035631489008665085,
+      "learning_rate": 0.00019733731226800015,
+      "loss": 0.7063,
+      "step": 274
+    },
+    {
+      "epoch": 0.15,
+      "grad_norm": 0.034868933260440826,
+      "learning_rate": 0.0001973171930399484,
+      "loss": 0.7211,
+      "step": 275
+    },
+    {
+      "epoch": 0.15,
+      "grad_norm": 0.030991340056061745,
+      "learning_rate": 0.00019729699912081979,
+      "loss": 0.7908,
+      "step": 276
+    },
+    {
+      "epoch": 0.15,
+      "grad_norm": 0.034880708903074265,
+      "learning_rate": 0.00019727673052611315,
+      "loss": 0.6827,
+      "step": 277
+    },
+    {
+      "epoch": 0.15,
+      "grad_norm": 0.03172425180673599,
+      "learning_rate": 0.00019725638727138468,
+      "loss": 0.7884,
+      "step": 278
+    },
+    {
+      "epoch": 0.16,
+      "grad_norm": 0.035629045218229294,
+      "learning_rate": 0.00019723596937224781,
+      "loss": 0.8101,
+      "step": 279
+    },
+    {
+      "epoch": 0.16,
+      "grad_norm": 0.03561673313379288,
+      "learning_rate": 0.00019721547684437333,
+      "loss": 0.8164,
+      "step": 280
+    },
+    {
+      "epoch": 0.16,
+      "grad_norm": 0.0325719490647316,
+      "learning_rate": 0.0001971949097034893,
+      "loss": 0.7356,
+      "step": 281
+    },
+    {
+      "epoch": 0.16,
+      "grad_norm": 0.04635859653353691,
+      "learning_rate": 0.00019717426796538102,
+      "loss": 0.9928,
+      "step": 282
+    },
+    {
+      "epoch": 0.16,
+      "grad_norm": 0.0328851044178009,
+      "learning_rate": 0.00019715355164589106,
+      "loss": 0.7742,
+      "step": 283
+    },
+    {
+      "epoch": 0.16,
+      "grad_norm": 0.032772939652204514,
+      "learning_rate": 0.00019713276076091917,
+      "loss": 0.7487,
+      "step": 284
+    },
+    {
+      "epoch": 0.16,
+      "grad_norm": 0.03649919852614403,
+      "learning_rate": 0.00019711189532642243,
+      "loss": 0.7322,
+      "step": 285
+    },
+    {
+      "epoch": 0.16,
+      "grad_norm": 0.03704332187771797,
+      "learning_rate": 0.0001970909553584151,
+      "loss": 0.7951,
+      "step": 286
+    },
+    {
+      "epoch": 0.16,
+      "grad_norm": 0.03600141033530235,
+      "learning_rate": 0.00019706994087296859,
+      "loss": 0.8053,
+      "step": 287
+    },
+    {
+      "epoch": 0.16,
+      "grad_norm": 0.03415602818131447,
+      "learning_rate": 0.0001970488518862116,
+      "loss": 0.8079,
+      "step": 288
+    },
+    {
+      "epoch": 0.16,
+      "grad_norm": 0.032455701380968094,
+      "learning_rate": 0.0001970276884143299,
+      "loss": 0.8268,
+      "step": 289
+    },
+    {
+      "epoch": 0.16,
+      "grad_norm": 0.03594019636511803,
+      "learning_rate": 0.00019700645047356652,
+      "loss": 0.7773,
+      "step": 290
+    },
+    {
+      "epoch": 0.16,
+      "grad_norm": 0.03449614718556404,
+      "learning_rate": 0.0001969851380802216,
+      "loss": 0.8421,
+      "step": 291
+    },
+    {
+      "epoch": 0.16,
+      "grad_norm": 0.035057246685028076,
+      "learning_rate": 0.0001969637512506524,
+      "loss": 0.7891,
+      "step": 292
+    },
+    {
+      "epoch": 0.16,
+      "grad_norm": 0.03100276179611683,
+      "learning_rate": 0.00019694229000127337,
+      "loss": 0.7188,
+      "step": 293
+    },
+    {
+      "epoch": 0.16,
+      "grad_norm": 0.035093460232019424,
+      "learning_rate": 0.00019692075434855606,
+      "loss": 0.6879,
+      "step": 294
+    },
+    {
+      "epoch": 0.16,
+      "grad_norm": 0.03476954251527786,
+      "learning_rate": 0.0001968991443090291,
+      "loss": 0.8428,
+      "step": 295
+    },
+    {
+      "epoch": 0.16,
+      "grad_norm": 0.03413707762956619,
+      "learning_rate": 0.00019687745989927823,
+      "loss": 0.8469,
+      "step": 296
+    },
+    {
+      "epoch": 0.17,
+      "grad_norm": 0.03546985611319542,
+      "learning_rate": 0.00019685570113594624,
+      "loss": 0.8675,
+      "step": 297
+    },
+    {
+      "epoch": 0.17,
+      "grad_norm": 0.036925192922353745,
+      "learning_rate": 0.00019683386803573302,
+      "loss": 0.7033,
+      "step": 298
+    },
+    {
+      "epoch": 0.17,
+      "grad_norm": 0.03350600227713585,
+      "learning_rate": 0.00019681196061539552,
+      "loss": 0.7405,
+      "step": 299
+    },
+    {
+      "epoch": 0.17,
+      "grad_norm": 0.03388972207903862,
+      "learning_rate": 0.00019678997889174772,
+      "loss": 0.7672,
+      "step": 300
+    },
+    {
+      "epoch": 0.17,
+      "grad_norm": 0.03225654736161232,
+      "learning_rate": 0.00019676792288166056,
+      "loss": 0.8161,
+      "step": 301
+    },
+    {
+      "epoch": 0.17,
+      "grad_norm": 0.035303112119436264,
+      "learning_rate": 0.00019674579260206215,
+      "loss": 0.8378,
+      "step": 302
+    },
+    {
+      "epoch": 0.17,
+      "grad_norm": 0.03430440276861191,
+      "learning_rate": 0.00019672358806993744,
+      "loss": 0.8755,
+      "step": 303
+    },
+    {
+      "epoch": 0.17,
+      "grad_norm": 0.03425530344247818,
+      "learning_rate": 0.00019670130930232849,
+      "loss": 0.8579,
+      "step": 304
+    },
+    {
+      "epoch": 0.17,
+      "grad_norm": 0.03408996760845184,
+      "learning_rate": 0.00019667895631633427,
+      "loss": 0.7241,
+      "step": 305
+    },
+    {
+      "epoch": 0.17,
+      "grad_norm": 0.034297868609428406,
+      "learning_rate": 0.00019665652912911067,
+      "loss": 0.7586,
+      "step": 306
+    },
+    {
+      "epoch": 0.17,
+      "grad_norm": 0.03555282577872276,
+      "learning_rate": 0.00019663402775787066,
+      "loss": 0.7913,
+      "step": 307
+    },
+    {
+      "epoch": 0.17,
+      "grad_norm": 0.046342890709638596,
+      "learning_rate": 0.000196611452219884,
+      "loss": 0.756,
+      "step": 308
+    },
+    {
+      "epoch": 0.17,
+      "grad_norm": 0.035010941326618195,
+      "learning_rate": 0.00019658880253247752,
+      "loss": 0.7984,
+      "step": 309
+    },
+    {
+      "epoch": 0.17,
+      "grad_norm": 0.034292541444301605,
+      "learning_rate": 0.00019656607871303486,
+      "loss": 0.7529,
+      "step": 310
+    },
+    {
+      "epoch": 0.17,
+      "grad_norm": 0.034864772111177444,
+      "learning_rate": 0.00019654328077899655,
+      "loss": 0.7659,
+      "step": 311
+    },
+    {
+      "epoch": 0.17,
+      "grad_norm": 0.03819848969578743,
+      "learning_rate": 0.0001965204087478601,
+      "loss": 0.8601,
+      "step": 312
+    },
+    {
+      "epoch": 0.17,
+      "grad_norm": 0.03443516790866852,
+      "learning_rate": 0.00019649746263717974,
+      "loss": 0.755,
+      "step": 313
+    },
+    {
+      "epoch": 0.17,
+      "grad_norm": 0.039128776639699936,
+      "learning_rate": 0.00019647444246456672,
+      "loss": 0.8283,
+      "step": 314
+    },
+    {
+      "epoch": 0.18,
+      "grad_norm": 0.03328540921211243,
+      "learning_rate": 0.000196451348247689,
+      "loss": 0.8621,
+      "step": 315
+    },
+    {
+      "epoch": 0.18,
+      "grad_norm": 0.03752969205379486,
+      "learning_rate": 0.00019642818000427145,
+      "loss": 0.6716,
+      "step": 316
+    },
+    {
+      "epoch": 0.18,
+      "grad_norm": 0.033460091799497604,
+      "learning_rate": 0.00019640493775209574,
+      "loss": 0.644,
+      "step": 317
+    },
+    {
+      "epoch": 0.18,
+      "grad_norm": 0.03952079266309738,
+      "learning_rate": 0.00019638162150900027,
+      "loss": 0.8444,
+      "step": 318
+    },
+    {
+      "epoch": 0.18,
+      "grad_norm": 0.03375870734453201,
+      "learning_rate": 0.00019635823129288038,
+      "loss": 0.7582,
+      "step": 319
+    },
+    {
+      "epoch": 0.18,
+      "grad_norm": 0.0385657362639904,
+      "learning_rate": 0.00019633476712168803,
+      "loss": 0.6819,
+      "step": 320
+    },
+    {
+      "epoch": 0.18,
+      "grad_norm": 0.03305663540959358,
+      "learning_rate": 0.000196311229013432,
+      "loss": 0.769,
+      "step": 321
+    },
+    {
+      "epoch": 0.18,
+      "grad_norm": 0.033982716500759125,
+      "learning_rate": 0.00019628761698617782,
+      "loss": 0.8634,
+      "step": 322
+    },
+    {
+      "epoch": 0.18,
+      "grad_norm": 0.035749975591897964,
+      "learning_rate": 0.00019626393105804776,
+      "loss": 0.7829,
+      "step": 323
+    },
+    {
+      "epoch": 0.18,
+      "grad_norm": 0.03398462384939194,
+      "learning_rate": 0.00019624017124722086,
+      "loss": 0.803,
+      "step": 324
+    },
+    {
+      "epoch": 0.18,
+      "grad_norm": 0.03670307993888855,
+      "learning_rate": 0.0001962163375719327,
+      "loss": 0.9054,
+      "step": 325
+    },
+    {
+      "epoch": 0.18,
+      "grad_norm": 0.034984368830919266,
+      "learning_rate": 0.00019619243005047576,
+      "loss": 0.7222,
+      "step": 326
+    },
+    {
+      "epoch": 0.18,
+      "grad_norm": 0.03537537157535553,
+      "learning_rate": 0.00019616844870119904,
+      "loss": 0.8563,
+      "step": 327
+    },
+    {
+      "epoch": 0.18,
+      "grad_norm": 0.03322187438607216,
+      "learning_rate": 0.00019614439354250824,
+      "loss": 0.8273,
+      "step": 328
+    },
+    {
+      "epoch": 0.18,
+      "grad_norm": 0.03316264599561691,
+      "learning_rate": 0.00019612026459286578,
+      "loss": 0.6768,
+      "step": 329
+    },
+    {
+      "epoch": 0.18,
+      "grad_norm": 0.03384651243686676,
+      "learning_rate": 0.00019609606187079065,
+      "loss": 0.7345,
+      "step": 330
+    },
+    {
+      "epoch": 0.18,
+      "grad_norm": 0.03392638638615608,
+      "learning_rate": 0.0001960717853948584,
+      "loss": 0.7833,
+      "step": 331
+    },
+    {
+      "epoch": 0.18,
+      "grad_norm": 0.03634137287735939,
+      "learning_rate": 0.00019604743518370133,
+      "loss": 0.8709,
+      "step": 332
+    },
+    {
+      "epoch": 0.19,
+      "grad_norm": 0.03778916969895363,
+      "learning_rate": 0.00019602301125600828,
+      "loss": 0.8933,
+      "step": 333
+    },
+    {
+      "epoch": 0.19,
+      "grad_norm": 0.04155290871858597,
+      "learning_rate": 0.0001959985136305246,
+      "loss": 0.9487,
+      "step": 334
+    },
+    {
+      "epoch": 0.19,
+      "grad_norm": 0.03602902963757515,
+      "learning_rate": 0.00019597394232605223,
+      "loss": 0.8372,
+      "step": 335
+    },
+    {
+      "epoch": 0.19,
+      "grad_norm": 0.033098701387643814,
+      "learning_rate": 0.00019594929736144976,
+      "loss": 0.8942,
+      "step": 336
+    },
+    {
+      "epoch": 0.19,
+      "grad_norm": 0.04318083077669144,
+      "learning_rate": 0.00019592457875563214,
+      "loss": 0.7353,
+      "step": 337
+    },
+    {
+      "epoch": 0.19,
+      "grad_norm": 0.035677891224622726,
+      "learning_rate": 0.000195899786527571,
+      "loss": 0.7229,
+      "step": 338
+    },
+    {
+      "epoch": 0.19,
+      "grad_norm": 0.03783939778804779,
+      "learning_rate": 0.0001958749206962944,
+      "loss": 0.7931,
+      "step": 339
+    },
+    {
+      "epoch": 0.19,
+      "grad_norm": 0.03574792295694351,
+      "learning_rate": 0.00019584998128088684,
+      "loss": 0.8139,
+      "step": 340
+    },
+    {
+      "epoch": 0.19,
+      "grad_norm": 0.042883213609457016,
+      "learning_rate": 0.00019582496830048947,
+      "loss": 0.8044,
+      "step": 341
+    },
+    {
+      "epoch": 0.19,
+      "grad_norm": 0.040097806602716446,
+      "learning_rate": 0.00019579988177429968,
+      "loss": 0.7571,
+      "step": 342
+    },
+    {
+      "epoch": 0.19,
+      "grad_norm": 0.033648643642663956,
+      "learning_rate": 0.00019577472172157144,
+      "loss": 0.771,
+      "step": 343
+    },
+    {
+      "epoch": 0.19,
+      "grad_norm": 0.04123647138476372,
+      "learning_rate": 0.00019574948816161513,
+      "loss": 0.8319,
+      "step": 344
+    },
+    {
+      "epoch": 0.19,
+      "grad_norm": 0.036461152136325836,
+      "learning_rate": 0.00019572418111379758,
+      "loss": 0.8421,
+      "step": 345
+    },
+    {
+      "epoch": 0.19,
+      "grad_norm": 0.034517571330070496,
+      "learning_rate": 0.00019569880059754194,
+      "loss": 0.9179,
+      "step": 346
+    },
+    {
+      "epoch": 0.19,
+      "grad_norm": 0.03678009659051895,
+      "learning_rate": 0.00019567334663232776,
+      "loss": 0.7756,
+      "step": 347
+    },
+    {
+      "epoch": 0.19,
+      "grad_norm": 0.038212236016988754,
+      "learning_rate": 0.00019564781923769108,
+      "loss": 0.7422,
+      "step": 348
+    },
+    {
+      "epoch": 0.19,
+      "grad_norm": 0.039440006017684937,
+      "learning_rate": 0.00019562221843322415,
+      "loss": 0.8191,
+      "step": 349
+    },
+    {
+      "epoch": 0.19,
+      "grad_norm": 0.0428134948015213,
+      "learning_rate": 0.0001955965442385756,
+      "loss": 0.8473,
+      "step": 350
+    },
+    {
+      "epoch": 0.2,
+      "grad_norm": 0.03534878045320511,
+      "learning_rate": 0.0001955707966734505,
+      "loss": 0.7492,
+      "step": 351
+    },
+    {
+      "epoch": 0.2,
+      "grad_norm": 0.03833247721195221,
+      "learning_rate": 0.00019554497575761004,
+      "loss": 0.7775,
+      "step": 352
+    },
+    {
+      "epoch": 0.2,
+      "grad_norm": 0.03430182486772537,
+      "learning_rate": 0.00019551908151087187,
+      "loss": 0.8519,
+      "step": 353
+    },
+    {
+      "epoch": 0.2,
+      "grad_norm": 0.03104781173169613,
+      "learning_rate": 0.00019549311395310984,
+      "loss": 0.7292,
+      "step": 354
+    },
+    {
+      "epoch": 0.2,
+      "grad_norm": 0.033952996134757996,
+      "learning_rate": 0.00019546707310425407,
+      "loss": 0.7592,
+      "step": 355
+    },
+    {
+      "epoch": 0.2,
+      "grad_norm": 0.03525162115693092,
+      "learning_rate": 0.00019544095898429097,
+      "loss": 0.7484,
+      "step": 356
+    },
+    {
+      "epoch": 0.2,
+      "grad_norm": 0.035235751420259476,
+      "learning_rate": 0.00019541477161326318,
+      "loss": 0.7206,
+      "step": 357
+    },
+    {
+      "epoch": 0.2,
+      "grad_norm": 0.038712091743946075,
+      "learning_rate": 0.00019538851101126947,
+      "loss": 0.8494,
+      "step": 358
+    },
+    {
+      "epoch": 0.2,
+      "grad_norm": 0.034723225980997086,
+      "learning_rate": 0.00019536217719846497,
+      "loss": 0.784,
+      "step": 359
+    },
+    {
+      "epoch": 0.2,
+      "grad_norm": 0.03219115734100342,
+      "learning_rate": 0.0001953357701950609,
+      "loss": 0.7373,
+      "step": 360
+    },
+    {
+      "epoch": 0.2,
+      "grad_norm": 0.037317853420972824,
+      "learning_rate": 0.00019530929002132462,
+      "loss": 0.7768,
+      "step": 361
+    },
+    {
+      "epoch": 0.2,
+      "grad_norm": 0.033880557864904404,
+      "learning_rate": 0.00019528273669757972,
+      "loss": 0.8359,
+      "step": 362
+    },
+    {
+      "epoch": 0.2,
+      "grad_norm": 0.037638477981090546,
+      "learning_rate": 0.00019525611024420595,
+      "loss": 0.7786,
+      "step": 363
+    },
+    {
+      "epoch": 0.2,
+      "grad_norm": 0.03666766360402107,
+      "learning_rate": 0.0001952294106816391,
+      "loss": 0.7666,
+      "step": 364
+    },
+    {
+      "epoch": 0.2,
+      "grad_norm": 0.03592545911669731,
+      "learning_rate": 0.00019520263803037116,
+      "loss": 0.8049,
+      "step": 365
+    },
+    {
+      "epoch": 0.2,
+      "grad_norm": 0.03477128967642784,
+      "learning_rate": 0.00019517579231095018,
+      "loss": 0.8243,
+      "step": 366
+    },
+    {
+      "epoch": 0.2,
+      "grad_norm": 0.03351607918739319,
+      "learning_rate": 0.00019514887354398023,
+      "loss": 0.6382,
+      "step": 367
+    },
+    {
+      "epoch": 0.2,
+      "grad_norm": 0.03531970828771591,
+      "learning_rate": 0.00019512188175012157,
+      "loss": 0.7363,
+      "step": 368
+    },
+    {
+      "epoch": 0.21,
+      "grad_norm": 0.03541957587003708,
+      "learning_rate": 0.00019509481695009035,
+      "loss": 0.7136,
+      "step": 369
+    },
+    {
+      "epoch": 0.21,
+      "grad_norm": 0.03538355231285095,
+      "learning_rate": 0.00019506767916465894,
+      "loss": 0.8702,
+      "step": 370
+    },
+    {
+      "epoch": 0.21,
+      "grad_norm": 0.036403317004442215,
+      "learning_rate": 0.00019504046841465556,
+      "loss": 0.773,
+      "step": 371
+    },
+    {
+      "epoch": 0.21,
+      "grad_norm": 0.050964437425136566,
+      "learning_rate": 0.0001950131847209645,
+      "loss": 0.65,
+      "step": 372
+    },
+    {
+      "epoch": 0.21,
+      "grad_norm": 0.033455755561590195,
+      "learning_rate": 0.0001949858281045261,
+      "loss": 0.8465,
+      "step": 373
+    },
+    {
+      "epoch": 0.21,
+      "grad_norm": 0.038496553897857666,
+      "learning_rate": 0.00019495839858633648,
+      "loss": 0.9064,
+      "step": 374
+    },
+    {
+      "epoch": 0.21,
+      "grad_norm": 0.046744123101234436,
+      "learning_rate": 0.00019493089618744796,
+      "loss": 0.7656,
+      "step": 375
+    },
+    {
+      "epoch": 0.21,
+      "grad_norm": 0.03650742024183273,
+      "learning_rate": 0.00019490332092896858,
+      "loss": 0.7547,
+      "step": 376
+    },
+    {
+      "epoch": 0.21,
+      "grad_norm": 0.03537287563085556,
+      "learning_rate": 0.00019487567283206242,
+      "loss": 0.7773,
+      "step": 377
+    },
+    {
+      "epoch": 0.21,
+      "grad_norm": 0.03856958448886871,
+      "learning_rate": 0.00019484795191794944,
+      "loss": 0.8212,
+      "step": 378
+    },
+    {
+      "epoch": 0.21,
+      "grad_norm": 0.03926542028784752,
+      "learning_rate": 0.0001948201582079055,
+      "loss": 0.7704,
+      "step": 379
+    },
+    {
+      "epoch": 0.21,
+      "grad_norm": 0.03413626179099083,
+      "learning_rate": 0.00019479229172326222,
+      "loss": 0.8243,
+      "step": 380
+    },
+    {
+      "epoch": 0.21,
+      "grad_norm": 0.036304738372564316,
+      "learning_rate": 0.00019476435248540726,
+      "loss": 0.7403,
+      "step": 381
+    },
+    {
+      "epoch": 0.21,
+      "grad_norm": 0.035184238106012344,
+      "learning_rate": 0.00019473634051578396,
+      "loss": 0.7825,
+      "step": 382
+    },
+    {
+      "epoch": 0.21,
+      "grad_norm": 0.0506867952644825,
+      "learning_rate": 0.00019470825583589157,
+      "loss": 0.951,
+      "step": 383
+    },
+    {
+      "epoch": 0.21,
+      "grad_norm": 0.034371040761470795,
+      "learning_rate": 0.00019468009846728513,
+      "loss": 0.6246,
+      "step": 384
+    },
+    {
+      "epoch": 0.21,
+      "grad_norm": 0.036034394055604935,
+      "learning_rate": 0.00019465186843157546,
+      "loss": 0.8194,
+      "step": 385
+    },
+    {
+      "epoch": 0.21,
+      "grad_norm": 0.03933184966444969,
+      "learning_rate": 0.00019462356575042914,
+      "loss": 0.6906,
+      "step": 386
+    },
+    {
+      "epoch": 0.22,
+      "grad_norm": 0.03522591292858124,
+      "learning_rate": 0.00019459519044556846,
+      "loss": 0.6198,
+      "step": 387
+    },
+    {
+      "epoch": 0.22,
+      "grad_norm": 0.033830877393484116,
+      "learning_rate": 0.00019456674253877162,
+      "loss": 0.7374,
+      "step": 388
+    },
+    {
+      "epoch": 0.22,
+      "grad_norm": 0.03900926932692528,
+      "learning_rate": 0.00019453822205187232,
+      "loss": 0.7763,
+      "step": 389
+    },
+    {
+      "epoch": 0.22,
+      "grad_norm": 0.035312309861183167,
+      "learning_rate": 0.00019450962900676016,
+      "loss": 0.6672,
+      "step": 390
+    },
+    {
+      "epoch": 0.22,
+      "grad_norm": 0.034244999289512634,
+      "learning_rate": 0.0001944809634253803,
+      "loss": 0.6852,
+      "step": 391
+    },
+    {
+      "epoch": 0.22,
+      "grad_norm": 0.03672676905989647,
+      "learning_rate": 0.00019445222532973356,
+      "loss": 0.7416,
+      "step": 392
+    },
+    {
+      "epoch": 0.22,
+      "grad_norm": 0.03532637655735016,
+      "learning_rate": 0.00019442341474187658,
+      "loss": 0.7647,
+      "step": 393
+    },
+    {
+      "epoch": 0.22,
+      "grad_norm": 0.03439967334270477,
+      "learning_rate": 0.00019439453168392144,
+      "loss": 0.7458,
+      "step": 394
+    },
+    {
+      "epoch": 0.22,
+      "grad_norm": 0.03463077172636986,
+      "learning_rate": 0.00019436557617803595,
+      "loss": 0.8125,
+      "step": 395
+    },
+    {
+      "epoch": 0.22,
+      "grad_norm": 0.03627005219459534,
+      "learning_rate": 0.00019433654824644347,
+      "loss": 0.6992,
+      "step": 396
+    },
+    {
+      "epoch": 0.22,
+      "grad_norm": 0.03494995832443237,
+      "learning_rate": 0.00019430744791142305,
+      "loss": 0.7078,
+      "step": 397
+    },
+    {
+      "epoch": 0.22,
+      "grad_norm": 0.03355133906006813,
+      "learning_rate": 0.00019427827519530918,
+      "loss": 0.6952,
+      "step": 398
+    },
+    {
+      "epoch": 0.22,
+      "grad_norm": 0.03713174909353256,
+      "learning_rate": 0.00019424903012049195,
+      "loss": 0.7782,
+      "step": 399
+    },
+    {
+      "epoch": 0.22,
+      "grad_norm": 0.03741822391748428,
+      "learning_rate": 0.00019421971270941701,
+      "loss": 0.7951,
+      "step": 400
+    },
+    {
+      "epoch": 0.22,
+      "grad_norm": 0.03539593145251274,
+      "learning_rate": 0.00019419032298458554,
+      "loss": 0.6305,
+      "step": 401
+    },
+    {
+      "epoch": 0.22,
+      "grad_norm": 0.03932331129908562,
+      "learning_rate": 0.00019416086096855418,
+      "loss": 0.6978,
+      "step": 402
+    },
+    {
+      "epoch": 0.22,
+      "grad_norm": 0.03865070268511772,
+      "learning_rate": 0.00019413132668393503,
+      "loss": 0.7058,
+      "step": 403
+    },
+    {
+      "epoch": 0.22,
+      "grad_norm": 0.036367420107126236,
+      "learning_rate": 0.00019410172015339576,
+      "loss": 0.7991,
+      "step": 404
+    },
+    {
+      "epoch": 0.23,
+      "grad_norm": 0.04757410287857056,
+      "learning_rate": 0.00019407204139965936,
+      "loss": 0.7277,
+      "step": 405
+    },
+    {
+      "epoch": 0.23,
+      "grad_norm": 0.04415794834494591,
+      "learning_rate": 0.00019404229044550433,
+      "loss": 0.7263,
+      "step": 406
+    },
+    {
+      "epoch": 0.23,
+      "grad_norm": 0.030960189178586006,
+      "learning_rate": 0.0001940124673137646,
+      "loss": 0.7514,
+      "step": 407
+    },
+    {
+      "epoch": 0.23,
+      "grad_norm": 0.0355895459651947,
+      "learning_rate": 0.0001939825720273294,
+      "loss": 0.9352,
+      "step": 408
+    },
+    {
+      "epoch": 0.23,
+      "grad_norm": 0.034076202660799026,
+      "learning_rate": 0.00019395260460914346,
+      "loss": 0.7354,
+      "step": 409
+    },
+    {
+      "epoch": 0.23,
+      "grad_norm": 0.03768068179488182,
+      "learning_rate": 0.00019392256508220682,
+      "loss": 0.797,
+      "step": 410
+    },
+    {
+      "epoch": 0.23,
+      "grad_norm": 0.03416946902871132,
+      "learning_rate": 0.00019389245346957478,
+      "loss": 0.8355,
+      "step": 411
+    },
+    {
+      "epoch": 0.23,
+      "grad_norm": 0.036176249384880066,
+      "learning_rate": 0.00019386226979435813,
+      "loss": 0.775,
+      "step": 412
+    },
+    {
+      "epoch": 0.23,
+      "grad_norm": 0.036431025713682175,
+      "learning_rate": 0.00019383201407972286,
+      "loss": 0.6291,
+      "step": 413
+    },
+    {
+      "epoch": 0.23,
+      "grad_norm": 0.03727782517671585,
+      "learning_rate": 0.00019380168634889025,
+      "loss": 0.8582,
+      "step": 414
+    },
+    {
+      "epoch": 0.23,
+      "grad_norm": 0.03564276918768883,
+      "learning_rate": 0.00019377128662513687,
+      "loss": 0.7397,
+      "step": 415
+    },
+    {
+      "epoch": 0.23,
+      "grad_norm": 0.037083700299263,
+      "learning_rate": 0.00019374081493179457,
+      "loss": 0.7173,
+      "step": 416
+    },
+    {
+      "epoch": 0.23,
+      "grad_norm": 0.040047336369752884,
+      "learning_rate": 0.00019371027129225042,
+      "loss": 0.6917,
+      "step": 417
+    },
+    {
+      "epoch": 0.23,
+      "grad_norm": 0.038617394864559174,
+      "learning_rate": 0.00019367965572994667,
+      "loss": 0.6494,
+      "step": 418
+    },
+    {
+      "epoch": 0.23,
+      "grad_norm": 0.03738728538155556,
+      "learning_rate": 0.00019364896826838083,
+      "loss": 0.8023,
+      "step": 419
+    },
+    {
+      "epoch": 0.23,
+      "grad_norm": 0.03948422148823738,
+      "learning_rate": 0.00019361820893110557,
+      "loss": 0.8126,
+      "step": 420
+    },
+    {
+      "epoch": 0.23,
+      "grad_norm": 0.035054076462984085,
+      "learning_rate": 0.0001935873777417287,
+      "loss": 0.7377,
+      "step": 421
+    },
+    {
+      "epoch": 0.23,
+      "grad_norm": 0.04039939120411873,
+      "learning_rate": 0.00019355647472391328,
+      "loss": 0.7804,
+      "step": 422
+    },
+    {
+      "epoch": 0.24,
+      "grad_norm": 0.034298498183488846,
+      "learning_rate": 0.00019352549990137732,
+      "loss": 0.8092,
+      "step": 423
+    },
+    {
+      "epoch": 0.24,
+      "grad_norm": 0.035063210874795914,
+      "learning_rate": 0.00019349445329789404,
+      "loss": 0.7162,
+      "step": 424
+    },
+    {
+      "epoch": 0.24,
+      "grad_norm": 0.03656880930066109,
+      "learning_rate": 0.00019346333493729182,
+      "loss": 0.7648,
+      "step": 425
+    },
+    {
+      "epoch": 0.24,
+      "grad_norm": 0.03690245375037193,
+      "learning_rate": 0.000193432144843454,
+      "loss": 0.7194,
+      "step": 426
+    },
+    {
+      "epoch": 0.24,
+      "grad_norm": 0.04461353272199631,
+      "learning_rate": 0.00019340088304031905,
+      "loss": 0.6901,
+      "step": 427
+    },
+    {
+      "epoch": 0.24,
+      "grad_norm": 0.03801785781979561,
+      "learning_rate": 0.0001933695495518804,
+      "loss": 0.7416,
+      "step": 428
+    },
+    {
+      "epoch": 0.24,
+      "grad_norm": 0.03857959061861038,
+      "learning_rate": 0.00019333814440218656,
+      "loss": 0.7467,
+      "step": 429
+    },
+    {
+      "epoch": 0.24,
+      "grad_norm": 0.03680335357785225,
+      "learning_rate": 0.00019330666761534104,
+      "loss": 0.7313,
+      "step": 430
+    },
+    {
+      "epoch": 0.24,
+      "grad_norm": 0.03958728164434433,
+      "learning_rate": 0.00019327511921550232,
+      "loss": 0.7054,
+      "step": 431
+    },
+    {
+      "epoch": 0.24,
+      "grad_norm": 0.03355373814702034,
+      "learning_rate": 0.0001932434992268838,
+      "loss": 0.7193,
+      "step": 432
+    },
+    {
+      "epoch": 0.24,
+      "grad_norm": 0.035141825675964355,
+      "learning_rate": 0.00019321180767375393,
+      "loss": 0.7841,
+      "step": 433
+    },
+    {
+      "epoch": 0.24,
+      "grad_norm": 0.037220679223537445,
+      "learning_rate": 0.00019318004458043596,
+      "loss": 0.7195,
+      "step": 434
+    },
+    {
+      "epoch": 0.24,
+      "grad_norm": 0.040347468107938766,
+      "learning_rate": 0.00019314820997130814,
+      "loss": 0.7725,
+      "step": 435
+    },
+    {
+      "epoch": 0.24,
+      "grad_norm": 0.03546297922730446,
+      "learning_rate": 0.00019311630387080356,
+      "loss": 0.7774,
+      "step": 436
+    },
+    {
+      "epoch": 0.24,
+      "grad_norm": 0.03774712607264519,
+      "learning_rate": 0.0001930843263034102,
+      "loss": 0.842,
+      "step": 437
+    },
+    {
+      "epoch": 0.24,
+      "grad_norm": 0.0372396744787693,
+      "learning_rate": 0.0001930522772936709,
+      "loss": 0.8061,
+      "step": 438
+    },
+    {
+      "epoch": 0.24,
+      "grad_norm": 0.03598784655332565,
+      "learning_rate": 0.00019302015686618326,
+      "loss": 0.8289,
+      "step": 439
+    },
+    {
+      "epoch": 0.24,
+      "grad_norm": 0.038949351757764816,
+      "learning_rate": 0.00019298796504559982,
+      "loss": 0.9173,
+      "step": 440
+    },
+    {
+      "epoch": 0.25,
+      "grad_norm": 0.03886549547314644,
+      "learning_rate": 0.00019295570185662783,
+      "loss": 0.7436,
+      "step": 441
+    },
+    {
+      "epoch": 0.25,
+      "grad_norm": 0.03816141560673714,
+      "learning_rate": 0.0001929233673240293,
+      "loss": 0.7178,
+      "step": 442
+    },
+    {
+      "epoch": 0.25,
+      "grad_norm": 0.03811522200703621,
+      "learning_rate": 0.00019289096147262106,
+      "loss": 0.7304,
+      "step": 443
+    },
+    {
+      "epoch": 0.25,
+      "grad_norm": 0.03597824648022652,
+      "learning_rate": 0.00019285848432727465,
+      "loss": 0.9174,
+      "step": 444
+    },
+    {
+      "epoch": 0.25,
+      "grad_norm": 0.036536138504743576,
+      "learning_rate": 0.0001928259359129163,
+      "loss": 0.6808,
+      "step": 445
+    },
+    {
+      "epoch": 0.25,
+      "grad_norm": 0.03995755687355995,
+      "learning_rate": 0.00019279331625452696,
+      "loss": 0.8131,
+      "step": 446
+    },
+    {
+      "epoch": 0.25,
+      "grad_norm": 0.03879891335964203,
+      "learning_rate": 0.00019276062537714226,
+      "loss": 0.7044,
+      "step": 447
+    },
+    {
+      "epoch": 0.25,
+      "grad_norm": 0.04052872955799103,
+      "learning_rate": 0.00019272786330585252,
+      "loss": 0.7703,
+      "step": 448
+    },
+    {
+      "epoch": 0.25,
+      "grad_norm": 0.047087304294109344,
+      "learning_rate": 0.00019269503006580266,
+      "loss": 0.764,
+      "step": 449
+    },
+    {
+      "epoch": 0.25,
+      "grad_norm": 0.035790085792541504,
+      "learning_rate": 0.0001926621256821922,
+      "loss": 0.72,
+      "step": 450
+    },
+    {
+      "epoch": 0.25,
+      "grad_norm": 0.03301222249865532,
+      "learning_rate": 0.00019262915018027538,
+      "loss": 0.8394,
+      "step": 451
+    },
+    {
+      "epoch": 0.25,
+      "grad_norm": 0.0383736714720726,
+      "learning_rate": 0.00019259610358536087,
+      "loss": 0.8389,
+      "step": 452
+    },
+    {
+      "epoch": 0.25,
+      "grad_norm": 0.0351652093231678,
+      "learning_rate": 0.00019256298592281202,
+      "loss": 0.7535,
+      "step": 453
+    },
+    {
+      "epoch": 0.25,
+      "grad_norm": 0.03904608264565468,
+      "learning_rate": 0.0001925297972180466,
+      "loss": 0.7091,
+      "step": 454
+    },
+    {
+      "epoch": 0.25,
+      "grad_norm": 0.035092540085315704,
+      "learning_rate": 0.00019249653749653711,
+      "loss": 0.6403,
+      "step": 455
+    },
+    {
+      "epoch": 0.25,
+      "grad_norm": 0.03878278285264969,
+      "learning_rate": 0.00019246320678381035,
+      "loss": 0.761,
+      "step": 456
+    },
+    {
+      "epoch": 0.25,
+      "grad_norm": 0.055852390825748444,
+      "learning_rate": 0.0001924298051054477,
+      "loss": 1.0766,
+      "step": 457
+    },
+    {
+      "epoch": 0.25,
+      "grad_norm": 0.038434069603681564,
+      "learning_rate": 0.000192396332487085,
+      "loss": 0.7273,
+      "step": 458
+    },
+    {
+      "epoch": 0.26,
+      "grad_norm": 0.04319579154253006,
+      "learning_rate": 0.0001923627889544125,
+      "loss": 0.7543,
+      "step": 459
+    },
+    {
+      "epoch": 0.26,
+      "grad_norm": 0.03686597943305969,
+      "learning_rate": 0.00019232917453317495,
+      "loss": 0.7785,
+      "step": 460
+    },
+    {
+      "epoch": 0.26,
+      "grad_norm": 0.03705907613039017,
+      "learning_rate": 0.00019229548924917146,
+      "loss": 0.8554,
+      "step": 461
+    },
+    {
+      "epoch": 0.26,
+      "grad_norm": 0.043490804731845856,
+      "learning_rate": 0.00019226173312825553,
+      "loss": 0.7574,
+      "step": 462
+    },
+    {
+      "epoch": 0.26,
+      "grad_norm": 0.0411020964384079,
+      "learning_rate": 0.00019222790619633499,
+      "loss": 0.7173,
+      "step": 463
+    },
+    {
+      "epoch": 0.26,
+      "grad_norm": 0.04355505481362343,
+      "learning_rate": 0.0001921940084793721,
+      "loss": 0.7586,
+      "step": 464
+    },
+    {
+      "epoch": 0.26,
+      "grad_norm": 0.03533725440502167,
+      "learning_rate": 0.0001921600400033834,
+      "loss": 0.7315,
+      "step": 465
+    },
+    {
+      "epoch": 0.26,
+      "grad_norm": 0.037127457559108734,
+      "learning_rate": 0.0001921260007944397,
+      "loss": 0.6851,
+      "step": 466
+    },
+    {
+      "epoch": 0.26,
+      "grad_norm": 0.04237948730587959,
+      "learning_rate": 0.00019209189087866614,
+      "loss": 0.7746,
+      "step": 467
+    },
+    {
+      "epoch": 0.26,
+      "grad_norm": 0.03694011643528938,
+      "learning_rate": 0.0001920577102822422,
+      "loss": 0.8222,
+      "step": 468
+    },
+    {
+      "epoch": 0.26,
+      "grad_norm": 0.03798561915755272,
+      "learning_rate": 0.00019202345903140147,
+      "loss": 0.7533,
+      "step": 469
+    },
+    {
+      "epoch": 0.26,
+      "grad_norm": 0.03925656899809837,
+      "learning_rate": 0.00019198913715243182,
+      "loss": 0.7317,
+      "step": 470
+    },
+    {
+      "epoch": 0.26,
+      "grad_norm": 0.03404255583882332,
+      "learning_rate": 0.00019195474467167544,
+      "loss": 0.7474,
+      "step": 471
+    },
+    {
+      "epoch": 0.26,
+      "grad_norm": 0.03907306119799614,
+      "learning_rate": 0.00019192028161552847,
+      "loss": 0.8662,
+      "step": 472
+    },
+    {
+      "epoch": 0.26,
+      "grad_norm": 0.040038101375103,
+      "learning_rate": 0.0001918857480104414,
+      "loss": 0.811,
+      "step": 473
+    },
+    {
+      "epoch": 0.26,
+      "grad_norm": 0.04068785905838013,
+      "learning_rate": 0.00019185114388291886,
+      "loss": 0.7831,
+      "step": 474
+    },
+    {
+      "epoch": 0.26,
+      "grad_norm": 0.03962669521570206,
+      "learning_rate": 0.00019181646925951954,
+      "loss": 0.7659,
+      "step": 475
+    },
+    {
+      "epoch": 0.26,
+      "grad_norm": 0.036200929433107376,
+      "learning_rate": 0.00019178172416685628,
+      "loss": 0.8251,
+      "step": 476
+    },
+    {
+      "epoch": 0.27,
+      "grad_norm": 0.03592352196574211,
+      "learning_rate": 0.00019174690863159593,
+      "loss": 0.7206,
+      "step": 477
+    },
+    {
+      "epoch": 0.27,
+      "grad_norm": 0.03576627001166344,
+      "learning_rate": 0.0001917120226804595,
+      "loss": 0.8105,
+      "step": 478
+    },
+    {
+      "epoch": 0.27,
+      "grad_norm": 0.034843627363443375,
+      "learning_rate": 0.00019167706634022199,
+      "loss": 0.8635,
+      "step": 479
+    },
+    {
+      "epoch": 0.27,
+      "grad_norm": 0.040766939520835876,
+      "learning_rate": 0.00019164203963771244,
+      "loss": 0.6993,
+      "step": 480
+    },
+    {
+      "epoch": 0.27,
+      "grad_norm": 0.04198118671774864,
+      "learning_rate": 0.00019160694259981388,
+      "loss": 0.8341,
+      "step": 481
+    },
+    {
+      "epoch": 0.27,
+      "grad_norm": 0.03662072494626045,
+      "learning_rate": 0.0001915717752534634,
+      "loss": 0.794,
+      "step": 482
+    },
+    {
+      "epoch": 0.27,
+      "grad_norm": 0.0427091047167778,
+      "learning_rate": 0.0001915365376256519,
+      "loss": 0.7304,
+      "step": 483
+    },
+    {
+      "epoch": 0.27,
+      "grad_norm": 0.03634573519229889,
+      "learning_rate": 0.00019150122974342435,
+      "loss": 0.7365,
+      "step": 484
+    },
+    {
+      "epoch": 0.27,
+      "grad_norm": 0.03949028626084328,
+      "learning_rate": 0.0001914658516338796,
+      "loss": 0.8231,
+      "step": 485
+    },
+    {
+      "epoch": 0.27,
+      "grad_norm": 0.04077625274658203,
+      "learning_rate": 0.00019143040332417037,
+      "loss": 0.66,
+      "step": 486
+    },
+    {
+      "epoch": 0.27,
+      "grad_norm": 0.03831211104989052,
+      "learning_rate": 0.0001913948848415033,
+      "loss": 0.8029,
+      "step": 487
+    },
+    {
+      "epoch": 0.27,
+      "grad_norm": 0.042038679122924805,
+      "learning_rate": 0.0001913592962131389,
+      "loss": 0.807,
+      "step": 488
+    },
+    {
+      "epoch": 0.27,
+      "grad_norm": 0.040266357362270355,
+      "learning_rate": 0.00019132363746639147,
+      "loss": 0.6887,
+      "step": 489
+    },
+    {
+      "epoch": 0.27,
+      "grad_norm": 0.0406985804438591,
+      "learning_rate": 0.0001912879086286291,
+      "loss": 0.7928,
+      "step": 490
+    },
+    {
+      "epoch": 0.27,
+      "grad_norm": 0.03616916015744209,
+      "learning_rate": 0.00019125210972727378,
+      "loss": 0.697,
+      "step": 491
+    },
+    {
+      "epoch": 0.27,
+      "grad_norm": 0.0417737253010273,
+      "learning_rate": 0.00019121624078980123,
+      "loss": 0.8405,
+      "step": 492
+    },
+    {
+      "epoch": 0.27,
+      "grad_norm": 0.04053768515586853,
+      "learning_rate": 0.00019118030184374086,
+      "loss": 0.7304,
+      "step": 493
+    },
+    {
+      "epoch": 0.27,
+      "grad_norm": 0.037591926753520966,
+      "learning_rate": 0.00019114429291667583,
+      "loss": 0.797,
+      "step": 494
+    },
+    {
+      "epoch": 0.28,
+      "grad_norm": 0.03530837222933769,
+      "learning_rate": 0.00019110821403624316,
+      "loss": 0.759,
+      "step": 495
+    },
+    {
+      "epoch": 0.28,
+      "grad_norm": 0.035193149000406265,
+      "learning_rate": 0.0001910720652301333,
+      "loss": 0.7179,
+      "step": 496
+    },
+    {
+      "epoch": 0.28,
+      "grad_norm": 0.03709745407104492,
+      "learning_rate": 0.0001910358465260906,
+      "loss": 0.8888,
+      "step": 497
+    },
+    {
+      "epoch": 0.28,
+      "grad_norm": 0.036165978759527206,
+      "learning_rate": 0.00019099955795191296,
+      "loss": 0.8409,
+      "step": 498
+    },
+    {
+      "epoch": 0.28,
+      "grad_norm": 0.03585299476981163,
+      "learning_rate": 0.00019096319953545185,
+      "loss": 0.6962,
+      "step": 499
+    },
+    {
+      "epoch": 0.28,
+      "grad_norm": 0.03879237174987793,
+      "learning_rate": 0.00019092677130461245,
+      "loss": 0.7037,
+      "step": 500
+    },
+    {
+      "epoch": 0.28,
+      "grad_norm": 0.040262021124362946,
+      "learning_rate": 0.00019089027328735348,
+      "loss": 0.8989,
+      "step": 501
+    },
+    {
+      "epoch": 0.28,
+      "grad_norm": 0.09765855222940445,
+      "learning_rate": 0.0001908537055116872,
+      "loss": 0.8365,
+      "step": 502
+    },
+    {
+      "epoch": 0.28,
+      "grad_norm": 0.04247608780860901,
+      "learning_rate": 0.00019081706800567943,
+      "loss": 0.8854,
+      "step": 503
+    },
+    {
+      "epoch": 0.28,
+      "grad_norm": 0.04151815176010132,
+      "learning_rate": 0.00019078036079744947,
+      "loss": 0.7774,
+      "step": 504
+    },
+    {
+      "epoch": 0.28,
+      "grad_norm": 0.04607046768069267,
+      "learning_rate": 0.00019074358391517023,
+      "loss": 0.7117,
+      "step": 505
+    },
+    {
+      "epoch": 0.28,
+      "grad_norm": 0.047151170670986176,
+      "learning_rate": 0.00019070673738706798,
+      "loss": 0.8029,
+      "step": 506
+    },
+    {
+      "epoch": 0.28,
+      "grad_norm": 0.0413680225610733,
+      "learning_rate": 0.00019066982124142244,
+      "loss": 0.6586,
+      "step": 507
+    },
+    {
+      "epoch": 0.28,
+      "grad_norm": 0.039332833141088486,
+      "learning_rate": 0.00019063283550656689,
+      "loss": 0.6051,
+      "step": 508
+    },
+    {
+      "epoch": 0.28,
+      "grad_norm": 0.03690175712108612,
+      "learning_rate": 0.0001905957802108878,
+      "loss": 0.6932,
+      "step": 509
+    },
+    {
+      "epoch": 0.28,
+      "grad_norm": 0.04148999601602554,
+      "learning_rate": 0.0001905586553828253,
+      "loss": 0.7663,
+      "step": 510
+    },
+    {
+      "epoch": 0.28,
+      "grad_norm": 0.03818729892373085,
+      "learning_rate": 0.00019052146105087267,
+      "loss": 0.7886,
+      "step": 511
+    },
+    {
+      "epoch": 0.28,
+      "grad_norm": 0.03793442249298096,
+      "learning_rate": 0.0001904841972435766,
+      "loss": 0.7099,
+      "step": 512
+    },
+    {
+      "epoch": 0.29,
+      "grad_norm": 0.04220060631632805,
+      "learning_rate": 0.00019044686398953715,
+      "loss": 0.7333,
+      "step": 513
+    },
+    {
+      "epoch": 0.29,
+      "grad_norm": 0.03924279659986496,
+      "learning_rate": 0.00019040946131740764,
+      "loss": 0.7619,
+      "step": 514
+    },
+    {
+      "epoch": 0.29,
+      "grad_norm": 0.03894231095910072,
+      "learning_rate": 0.00019037198925589465,
+      "loss": 0.8214,
+      "step": 515
+    },
+    {
+      "epoch": 0.29,
+      "grad_norm": 0.038151923567056656,
+      "learning_rate": 0.00019033444783375804,
+      "loss": 0.8511,
+      "step": 516
+    },
+    {
+      "epoch": 0.29,
+      "grad_norm": 0.04167793318629265,
+      "learning_rate": 0.0001902968370798109,
+      "loss": 0.7582,
+      "step": 517
+    },
+    {
+      "epoch": 0.29,
+      "grad_norm": 0.039893049746751785,
+      "learning_rate": 0.00019025915702291956,
+      "loss": 0.8991,
+      "step": 518
+    },
+    {
+      "epoch": 0.29,
+      "grad_norm": 0.03804799169301987,
+      "learning_rate": 0.00019022140769200352,
+      "loss": 0.7959,
+      "step": 519
+    },
+    {
+      "epoch": 0.29,
+      "grad_norm": 0.03987705335021019,
+      "learning_rate": 0.00019018358911603538,
+      "loss": 0.7415,
+      "step": 520
+    },
+    {
+      "epoch": 0.29,
+      "grad_norm": 0.036575764417648315,
+      "learning_rate": 0.000190145701324041,
+      "loss": 0.6772,
+      "step": 521
+    },
+    {
+      "epoch": 0.29,
+      "grad_norm": 0.03801872953772545,
+      "learning_rate": 0.00019010774434509933,
+      "loss": 0.7397,
+      "step": 522
+    },
+    {
+      "epoch": 0.29,
+      "grad_norm": 0.03547314926981926,
+      "learning_rate": 0.0001900697182083423,
+      "loss": 0.6802,
+      "step": 523
+    },
+    {
+      "epoch": 0.29,
+      "grad_norm": 0.03981153294444084,
+      "learning_rate": 0.00019003162294295515,
+      "loss": 0.8714,
+      "step": 524
+    },
+    {
+      "epoch": 0.29,
+      "grad_norm": 0.035997990518808365,
+      "learning_rate": 0.000189993458578176,
+      "loss": 0.6773,
+      "step": 525
+    },
+    {
+      "epoch": 0.29,
+      "grad_norm": 0.04118787497282028,
+      "learning_rate": 0.00018995522514329602,
+      "loss": 0.7247,
+      "step": 526
+    },
+    {
+      "epoch": 0.29,
+      "grad_norm": 0.04163433611392975,
+      "learning_rate": 0.00018991692266765947,
+      "loss": 0.7977,
+      "step": 527
+    },
+    {
+      "epoch": 0.29,
+      "grad_norm": 0.03958054631948471,
+      "learning_rate": 0.0001898785511806635,
+      "loss": 0.7665,
+      "step": 528
+    },
+    {
+      "epoch": 0.29,
+      "grad_norm": 0.040298108011484146,
+      "learning_rate": 0.0001898401107117583,
+      "loss": 0.8497,
+      "step": 529
+    },
+    {
+      "epoch": 0.29,
+      "grad_norm": 0.039216410368680954,
+      "learning_rate": 0.00018980160129044698,
+      "loss": 0.8379,
+      "step": 530
+    },
+    {
+      "epoch": 0.3,
+      "grad_norm": 0.037412021309137344,
+      "learning_rate": 0.0001897630229462856,
+      "loss": 0.8241,
+      "step": 531
+    },
+    {
+      "epoch": 0.3,
+      "grad_norm": 0.042515527456998825,
+      "learning_rate": 0.0001897243757088831,
+      "loss": 0.7151,
+      "step": 532
+    },
+    {
+      "epoch": 0.3,
+      "grad_norm": 0.038337383419275284,
+      "learning_rate": 0.0001896856596079012,
+      "loss": 0.7465,
+      "step": 533
+    },
+    {
+      "epoch": 0.3,
+      "grad_norm": 0.03905492648482323,
+      "learning_rate": 0.00018964687467305465,
+      "loss": 0.6749,
+      "step": 534
+    },
+    {
+      "epoch": 0.3,
+      "grad_norm": 0.038022320717573166,
+      "learning_rate": 0.0001896080209341109,
+      "loss": 0.6692,
+      "step": 535
+    },
+    {
+      "epoch": 0.3,
+      "grad_norm": 0.0380224771797657,
+      "learning_rate": 0.00018956909842089023,
+      "loss": 0.866,
+      "step": 536
+    },
+    {
+      "epoch": 0.3,
+      "grad_norm": 0.03461451828479767,
+      "learning_rate": 0.00018953010716326577,
+      "loss": 0.7494,
+      "step": 537
+    },
+    {
+      "epoch": 0.3,
+      "grad_norm": 0.04129749909043312,
+      "learning_rate": 0.00018949104719116332,
+      "loss": 0.783,
+      "step": 538
+    },
+    {
+      "epoch": 0.3,
+      "grad_norm": 0.0383567251265049,
+      "learning_rate": 0.00018945191853456152,
+      "loss": 0.7299,
+      "step": 539
+    },
+    {
+      "epoch": 0.3,
+      "grad_norm": 0.038509126752614975,
+      "learning_rate": 0.0001894127212234916,
+      "loss": 0.8074,
+      "step": 540
+    },
+    {
+      "epoch": 0.3,
+      "grad_norm": 0.0398080088198185,
+      "learning_rate": 0.00018937345528803755,
+      "loss": 0.7684,
+      "step": 541
+    },
+    {
+      "epoch": 0.3,
+      "grad_norm": 0.04023388773202896,
+      "learning_rate": 0.0001893341207583361,
+      "loss": 0.8234,
+      "step": 542
+    },
+    {
+      "epoch": 0.3,
+      "grad_norm": 0.0396123044192791,
+      "learning_rate": 0.00018929471766457648,
+      "loss": 0.8573,
+      "step": 543
+    },
+    {
+      "epoch": 0.3,
+      "grad_norm": 0.03965805470943451,
+      "learning_rate": 0.00018925524603700063,
+      "loss": 0.8654,
+      "step": 544
+    },
+    {
+      "epoch": 0.3,
+      "grad_norm": 0.03725765272974968,
+      "learning_rate": 0.00018921570590590316,
+      "loss": 0.7674,
+      "step": 545
+    },
+    {
+      "epoch": 0.3,
+      "grad_norm": 0.042921282351017,
+      "learning_rate": 0.00018917609730163105,
+      "loss": 0.7293,
+      "step": 546
+    },
+    {
+      "epoch": 0.3,
+      "grad_norm": 0.04281310364603996,
+      "learning_rate": 0.00018913642025458405,
+      "loss": 0.8283,
+      "step": 547
+    },
+    {
+      "epoch": 0.3,
+      "grad_norm": 0.03923151642084122,
+      "learning_rate": 0.00018909667479521426,
+      "loss": 0.8305,
+      "step": 548
+    },
+    {
+      "epoch": 0.31,
+      "grad_norm": 0.035802267491817474,
+      "learning_rate": 0.00018905686095402647,
+      "loss": 0.6796,
+      "step": 549
+    },
+    {
+      "epoch": 0.31,
+      "grad_norm": 0.04895975813269615,
+      "learning_rate": 0.00018901697876157777,
+      "loss": 0.8961,
+      "step": 550
+    },
+    {
+      "epoch": 0.31,
+      "grad_norm": 0.039671383798122406,
+      "learning_rate": 0.00018897702824847786,
+      "loss": 0.7407,
+      "step": 551
+    },
+    {
+      "epoch": 0.31,
+      "grad_norm": 0.038863249123096466,
+      "learning_rate": 0.00018893700944538883,
+      "loss": 0.7412,
+      "step": 552
+    },
+    {
+      "epoch": 0.31,
+      "grad_norm": 0.03964029625058174,
+      "learning_rate": 0.00018889692238302508,
+      "loss": 0.7407,
+      "step": 553
+    },
+    {
+      "epoch": 0.31,
+      "grad_norm": 0.03963325545191765,
+      "learning_rate": 0.00018885676709215355,
+      "loss": 0.671,
+      "step": 554
+    },
+    {
+      "epoch": 0.31,
+      "grad_norm": 0.04223211482167244,
+      "learning_rate": 0.00018881654360359344,
+      "loss": 0.9177,
+      "step": 555
+    },
+    {
+      "epoch": 0.31,
+      "grad_norm": 0.04407535493373871,
+      "learning_rate": 0.00018877625194821637,
+      "loss": 0.8018,
+      "step": 556
+    },
+    {
+      "epoch": 0.31,
+      "grad_norm": 0.045421354472637177,
+      "learning_rate": 0.00018873589215694623,
+      "loss": 0.7992,
+      "step": 557
+    },
+    {
+      "epoch": 0.31,
+      "grad_norm": 0.03759824484586716,
+      "learning_rate": 0.00018869546426075919,
+      "loss": 0.6757,
+      "step": 558
+    },
+    {
+      "epoch": 0.31,
+      "grad_norm": 0.03787359595298767,
+      "learning_rate": 0.00018865496829068373,
+      "loss": 0.7533,
+      "step": 559
+    },
+    {
+      "epoch": 0.31,
+      "grad_norm": 0.04168887808918953,
+      "learning_rate": 0.0001886144042778006,
+      "loss": 0.8529,
+      "step": 560
+    },
+    {
+      "epoch": 0.31,
+      "grad_norm": 0.03604380786418915,
+      "learning_rate": 0.00018857377225324265,
+      "loss": 0.6959,
+      "step": 561
+    },
+    {
+      "epoch": 0.31,
+      "grad_norm": 0.039811234921216965,
+      "learning_rate": 0.00018853307224819506,
+      "loss": 0.6923,
+      "step": 562
+    },
+    {
+      "epoch": 0.31,
+      "grad_norm": 0.039177920669317245,
+      "learning_rate": 0.00018849230429389516,
+      "loss": 0.8089,
+      "step": 563
+    },
+    {
+      "epoch": 0.31,
+      "grad_norm": 0.03779724985361099,
+      "learning_rate": 0.0001884514684216324,
+      "loss": 0.6745,
+      "step": 564
+    },
+    {
+      "epoch": 0.31,
+      "grad_norm": 0.03678756207227707,
+      "learning_rate": 0.00018841056466274836,
+      "loss": 0.6761,
+      "step": 565
+    },
+    {
+      "epoch": 0.31,
+      "grad_norm": 0.03539562225341797,
+      "learning_rate": 0.0001883695930486367,
+      "loss": 0.7111,
+      "step": 566
+    },
+    {
+      "epoch": 0.32,
+      "grad_norm": 0.03760172799229622,
+      "learning_rate": 0.00018832855361074322,
+      "loss": 0.8614,
+      "step": 567
+    },
+    {
+      "epoch": 0.32,
+      "grad_norm": 0.03915516659617424,
+      "learning_rate": 0.00018828744638056574,
+      "loss": 0.8167,
+      "step": 568
+    },
+    {
+      "epoch": 0.32,
+      "grad_norm": 0.039351411163806915,
+      "learning_rate": 0.00018824627138965414,
+      "loss": 0.7736,
+      "step": 569
+    },
+    {
+      "epoch": 0.32,
+      "grad_norm": 0.03440462425351143,
+      "learning_rate": 0.00018820502866961022,
+      "loss": 0.6989,
+      "step": 570
+    },
+    {
+      "epoch": 0.32,
+      "grad_norm": 0.04510714113712311,
+      "learning_rate": 0.00018816371825208789,
+      "loss": 0.7669,
+      "step": 571
+    },
+    {
+      "epoch": 0.32,
+      "grad_norm": 0.038387760519981384,
+      "learning_rate": 0.0001881223401687929,
+      "loss": 0.7742,
+      "step": 572
+    },
+    {
+      "epoch": 0.32,
+      "grad_norm": 0.04283255338668823,
+      "learning_rate": 0.00018808089445148302,
+      "loss": 0.8729,
+      "step": 573
+    },
+    {
+      "epoch": 0.32,
+      "grad_norm": 0.03683461621403694,
+      "learning_rate": 0.00018803938113196787,
+      "loss": 0.7528,
+      "step": 574
+    },
+    {
+      "epoch": 0.32,
+      "grad_norm": 0.03634823486208916,
+      "learning_rate": 0.00018799780024210898,
+      "loss": 0.7812,
+      "step": 575
+    },
+    {
+      "epoch": 0.32,
+      "grad_norm": 0.04318126663565636,
+      "learning_rate": 0.00018795615181381976,
+      "loss": 0.8744,
+      "step": 576
+    },
+    {
+      "epoch": 0.32,
+      "grad_norm": 0.04043075069785118,
+      "learning_rate": 0.00018791443587906542,
+      "loss": 0.719,
+      "step": 577
+    },
+    {
+      "epoch": 0.32,
+      "grad_norm": 0.03615717217326164,
+      "learning_rate": 0.00018787265246986298,
+      "loss": 0.6402,
+      "step": 578
+    },
+    {
+      "epoch": 0.32,
+      "grad_norm": 0.0358918234705925,
+      "learning_rate": 0.0001878308016182813,
+      "loss": 0.7614,
+      "step": 579
+    },
+    {
+      "epoch": 0.32,
+      "grad_norm": 0.03841692954301834,
+      "learning_rate": 0.0001877888833564409,
+      "loss": 0.6891,
+      "step": 580
+    },
+    {
+      "epoch": 0.32,
+      "grad_norm": 0.04361288622021675,
+      "learning_rate": 0.00018774689771651422,
+      "loss": 0.866,
+      "step": 581
+    },
+    {
+      "epoch": 0.32,
+      "grad_norm": 0.04058700054883957,
+      "learning_rate": 0.0001877048447307252,
+      "loss": 0.7412,
+      "step": 582
+    },
+    {
+      "epoch": 0.32,
+      "grad_norm": 0.04075286164879799,
+      "learning_rate": 0.00018766272443134955,
+      "loss": 0.8213,
+      "step": 583
+    },
+    {
+      "epoch": 0.32,
+      "grad_norm": 0.03992673009634018,
+      "learning_rate": 0.00018762053685071473,
+      "loss": 0.7053,
+      "step": 584
+    },
+    {
+      "epoch": 0.33,
+      "grad_norm": 0.03777340427041054,
+      "learning_rate": 0.00018757828202119973,
+      "loss": 0.656,
+      "step": 585
+    },
+    {
+      "epoch": 0.33,
+      "grad_norm": 0.04056774452328682,
+      "learning_rate": 0.00018753595997523514,
+      "loss": 0.8583,
+      "step": 586
+    },
+    {
+      "epoch": 0.33,
+      "grad_norm": 0.03643214702606201,
+      "learning_rate": 0.00018749357074530326,
+      "loss": 0.7368,
+      "step": 587
+    },
+    {
+      "epoch": 0.33,
+      "grad_norm": 0.040921106934547424,
+      "learning_rate": 0.00018745111436393788,
+      "loss": 0.7502,
+      "step": 588
+    },
+    {
+      "epoch": 0.33,
+      "grad_norm": 0.037460386753082275,
+      "learning_rate": 0.00018740859086372429,
+      "loss": 0.7211,
+      "step": 589
+    },
+    {
+      "epoch": 0.33,
+      "grad_norm": 0.04132199287414551,
+      "learning_rate": 0.00018736600027729932,
+      "loss": 0.8679,
+      "step": 590
+    },
+    {
+      "epoch": 0.33,
+      "grad_norm": 0.038056958466768265,
+      "learning_rate": 0.00018732334263735136,
+      "loss": 0.7041,
+      "step": 591
+    },
+    {
+      "epoch": 0.33,
+      "grad_norm": 0.036732178181409836,
+      "learning_rate": 0.0001872806179766202,
+      "loss": 0.8356,
+      "step": 592
+    },
+    {
+      "epoch": 0.33,
+      "grad_norm": 0.04578113928437233,
+      "learning_rate": 0.00018723782632789701,
+      "loss": 0.7287,
+      "step": 593
+    },
+    {
+      "epoch": 0.33,
+      "grad_norm": 0.04152052104473114,
+      "learning_rate": 0.0001871949677240245,
+      "loss": 0.7445,
+      "step": 594
+    },
+    {
+      "epoch": 0.33,
+      "grad_norm": 0.05432068929076195,
+      "learning_rate": 0.00018715204219789668,
+      "loss": 0.6701,
+      "step": 595
+    },
+    {
+      "epoch": 0.33,
+      "grad_norm": 0.040836673229932785,
+      "learning_rate": 0.00018710904978245894,
+      "loss": 0.7228,
+      "step": 596
+    },
+    {
+      "epoch": 0.33,
+      "grad_norm": 0.04091890901327133,
+      "learning_rate": 0.00018706599051070808,
+      "loss": 0.7615,
+      "step": 597
+    },
+    {
+      "epoch": 0.33,
+      "grad_norm": 0.039113231003284454,
+      "learning_rate": 0.00018702286441569206,
+      "loss": 0.8016,
+      "step": 598
+    },
+    {
+      "epoch": 0.33,
+      "grad_norm": 0.03947889804840088,
+      "learning_rate": 0.00018697967153051028,
+      "loss": 0.8327,
+      "step": 599
+    },
+    {
+      "epoch": 0.33,
+      "grad_norm": 0.03761962801218033,
+      "learning_rate": 0.0001869364118883133,
+      "loss": 0.7843,
+      "step": 600
+    },
+    {
+      "epoch": 0.33,
+      "grad_norm": 0.035844236612319946,
+      "learning_rate": 0.00018689308552230296,
+      "loss": 0.7678,
+      "step": 601
+    },
+    {
+      "epoch": 0.33,
+      "grad_norm": 0.04012085869908333,
+      "learning_rate": 0.00018684969246573235,
+      "loss": 0.8195,
+      "step": 602
+    },
+    {
+      "epoch": 0.34,
+      "grad_norm": 0.04106171056628227,
+      "learning_rate": 0.00018680623275190564,
+      "loss": 0.8068,
+      "step": 603
+    },
+    {
+      "epoch": 0.34,
+      "grad_norm": 0.03810250759124756,
+      "learning_rate": 0.00018676270641417822,
+      "loss": 0.6382,
+      "step": 604
+    },
+    {
+      "epoch": 0.34,
+      "grad_norm": 0.04322788119316101,
+      "learning_rate": 0.00018671911348595667,
+      "loss": 0.7498,
+      "step": 605
+    },
+    {
+      "epoch": 0.34,
+      "grad_norm": 0.04026930034160614,
+      "learning_rate": 0.00018667545400069858,
+      "loss": 0.7506,
+      "step": 606
+    },
+    {
+      "epoch": 0.34,
+      "grad_norm": 0.040973201394081116,
+      "learning_rate": 0.00018663172799191264,
+      "loss": 0.7408,
+      "step": 607
+    },
+    {
+      "epoch": 0.34,
+      "grad_norm": 0.03879684954881668,
+      "learning_rate": 0.00018658793549315868,
+      "loss": 0.805,
+      "step": 608
+    },
+    {
+      "epoch": 0.34,
+      "grad_norm": 0.038516003638505936,
+      "learning_rate": 0.00018654407653804746,
+      "loss": 0.6352,
+      "step": 609
+    },
+    {
+      "epoch": 0.34,
+      "grad_norm": 0.041547078639268875,
+      "learning_rate": 0.0001865001511602408,
+      "loss": 0.7504,
+      "step": 610
+    },
+    {
+      "epoch": 0.34,
+      "grad_norm": 0.03565856069326401,
+      "learning_rate": 0.0001864561593934515,
+      "loss": 0.6952,
+      "step": 611
+    },
+    {
+      "epoch": 0.34,
+      "grad_norm": 0.043429479002952576,
+      "learning_rate": 0.00018641210127144327,
+      "loss": 0.7584,
+      "step": 612
+    },
+    {
+      "epoch": 0.34,
+      "grad_norm": 0.043735940009355545,
+      "learning_rate": 0.00018636797682803082,
+      "loss": 0.6309,
+      "step": 613
+    },
+    {
+      "epoch": 0.34,
+      "grad_norm": 0.03792436793446541,
+      "learning_rate": 0.00018632378609707967,
+      "loss": 0.6673,
+      "step": 614
+    },
+    {
+      "epoch": 0.34,
+      "grad_norm": 0.03878699988126755,
+      "learning_rate": 0.0001862795291125063,
+      "loss": 0.6326,
+      "step": 615
+    },
+    {
+      "epoch": 0.34,
+      "grad_norm": 0.04062122479081154,
+      "learning_rate": 0.00018623520590827799,
+      "loss": 0.6807,
+      "step": 616
+    },
+    {
+      "epoch": 0.34,
+      "grad_norm": 0.046398624777793884,
+      "learning_rate": 0.00018619081651841286,
+      "loss": 0.6643,
+      "step": 617
+    },
+    {
+      "epoch": 0.34,
+      "grad_norm": 0.04108712822198868,
+      "learning_rate": 0.00018614636097697985,
+      "loss": 0.7254,
+      "step": 618
+    },
+    {
+      "epoch": 0.34,
+      "grad_norm": 0.0392933115363121,
+      "learning_rate": 0.00018610183931809862,
+      "loss": 0.7912,
+      "step": 619
+    },
+    {
+      "epoch": 0.34,
+      "grad_norm": 0.03968581184744835,
+      "learning_rate": 0.00018605725157593958,
+      "loss": 0.7038,
+      "step": 620
+    },
+    {
+      "epoch": 0.35,
+      "grad_norm": 0.03825043514370918,
+      "learning_rate": 0.00018601259778472395,
+      "loss": 0.6573,
+      "step": 621
+    },
+    {
+      "epoch": 0.35,
+      "grad_norm": 0.04096253588795662,
+      "learning_rate": 0.00018596787797872354,
+      "loss": 0.7289,
+      "step": 622
+    },
+    {
+      "epoch": 0.35,
+      "grad_norm": 0.03817110136151314,
+      "learning_rate": 0.00018592309219226083,
+      "loss": 0.7687,
+      "step": 623
+    },
+    {
+      "epoch": 0.35,
+      "grad_norm": 0.0371723398566246,
+      "learning_rate": 0.00018587824045970903,
+      "loss": 0.631,
+      "step": 624
+    },
+    {
+      "epoch": 0.35,
+      "grad_norm": 0.04173294082283974,
+      "learning_rate": 0.0001858333228154919,
+      "loss": 0.7503,
+      "step": 625
+    },
+    {
+      "epoch": 0.35,
+      "grad_norm": 0.03969898074865341,
+      "learning_rate": 0.0001857883392940837,
+      "loss": 0.7999,
+      "step": 626
+    },
+    {
+      "epoch": 0.35,
+      "grad_norm": 0.034285057336091995,
+      "learning_rate": 0.00018574328993000946,
+      "loss": 0.742,
+      "step": 627
+    },
+    {
+      "epoch": 0.35,
+      "grad_norm": 0.03918811306357384,
+      "learning_rate": 0.0001856981747578446,
+      "loss": 0.647,
+      "step": 628
+    },
+    {
+      "epoch": 0.35,
+      "grad_norm": 0.039601150900125504,
+      "learning_rate": 0.00018565299381221505,
+      "loss": 0.6368,
+      "step": 629
+    },
+    {
+      "epoch": 0.35,
+      "grad_norm": 0.04142673686146736,
+      "learning_rate": 0.0001856077471277972,
+      "loss": 0.8081,
+      "step": 630
+    },
+    {
+      "epoch": 0.35,
+      "grad_norm": 0.04136984422802925,
+      "learning_rate": 0.00018556243473931801,
+      "loss": 0.7583,
+      "step": 631
+    },
+    {
+      "epoch": 0.35,
+      "grad_norm": 0.046918828040361404,
+      "learning_rate": 0.0001855170566815548,
+      "loss": 0.8415,
+      "step": 632
+    },
+    {
+      "epoch": 0.35,
+      "grad_norm": 0.03764290362596512,
+      "learning_rate": 0.0001854716129893353,
+      "loss": 0.6575,
+      "step": 633
+    },
+    {
+      "epoch": 0.35,
+      "grad_norm": 0.04044364020228386,
+      "learning_rate": 0.00018542610369753755,
+      "loss": 0.6595,
+      "step": 634
+    },
+    {
+      "epoch": 0.35,
+      "grad_norm": 0.04079195857048035,
+      "learning_rate": 0.00018538052884109005,
+      "loss": 0.6936,
+      "step": 635
+    },
+    {
+      "epoch": 0.35,
+      "grad_norm": 0.04137120023369789,
+      "learning_rate": 0.00018533488845497148,
+      "loss": 0.7419,
+      "step": 636
+    },
+    {
+      "epoch": 0.35,
+      "grad_norm": 0.0409332811832428,
+      "learning_rate": 0.000185289182574211,
+      "loss": 0.7343,
+      "step": 637
+    },
+    {
+      "epoch": 0.35,
+      "grad_norm": 0.03855755180120468,
+      "learning_rate": 0.0001852434112338879,
+      "loss": 0.7341,
+      "step": 638
+    },
+    {
+      "epoch": 0.36,
+      "grad_norm": 0.04691479727625847,
+      "learning_rate": 0.0001851975744691317,
+      "loss": 0.6813,
+      "step": 639
+    },
+    {
+      "epoch": 0.36,
+      "grad_norm": 0.04192007705569267,
+      "learning_rate": 0.00018515167231512224,
+      "loss": 0.8011,
+      "step": 640
+    },
+    {
+      "epoch": 0.36,
+      "grad_norm": 0.04521464928984642,
+      "learning_rate": 0.00018510570480708942,
+      "loss": 0.7823,
+      "step": 641
+    },
+    {
+      "epoch": 0.36,
+      "grad_norm": 0.035000029951334,
+      "learning_rate": 0.0001850596719803134,
+      "loss": 0.5621,
+      "step": 642
+    },
+    {
+      "epoch": 0.36,
+      "grad_norm": 0.04110797494649887,
+      "learning_rate": 0.00018501357387012447,
+      "loss": 0.6851,
+      "step": 643
+    },
+    {
+      "epoch": 0.36,
+      "grad_norm": 0.04273466393351555,
+      "learning_rate": 0.00018496741051190297,
+      "loss": 0.7967,
+      "step": 644
+    },
+    {
+      "epoch": 0.36,
+      "grad_norm": 0.04418378323316574,
+      "learning_rate": 0.0001849211819410793,
+      "loss": 0.7684,
+      "step": 645
+    },
+    {
+      "epoch": 0.36,
+      "grad_norm": 0.03610142320394516,
+      "learning_rate": 0.00018487488819313402,
+      "loss": 0.6515,
+      "step": 646
+    },
+    {
+      "epoch": 0.36,
+      "grad_norm": 0.041490521281957626,
+      "learning_rate": 0.0001848285293035976,
+      "loss": 0.8379,
+      "step": 647
+    },
+    {
+      "epoch": 0.36,
+      "grad_norm": 0.041976895183324814,
+      "learning_rate": 0.0001847821053080505,
+      "loss": 0.7772,
+      "step": 648
+    },
+    {
+      "epoch": 0.36,
+      "grad_norm": 0.041484154760837555,
+      "learning_rate": 0.0001847356162421233,
+      "loss": 0.7797,
+      "step": 649
+    },
+    {
+      "epoch": 0.36,
+      "grad_norm": 0.04266678914427757,
+      "learning_rate": 0.00018468906214149638,
+      "loss": 0.681,
+      "step": 650
+    },
+    {
+      "epoch": 0.36,
+      "grad_norm": 0.041691817343235016,
+      "learning_rate": 0.00018464244304190004,
+      "loss": 0.8475,
+      "step": 651
+    },
+    {
+      "epoch": 0.36,
+      "grad_norm": 0.039543136954307556,
+      "learning_rate": 0.00018459575897911455,
+      "loss": 0.5914,
+      "step": 652
+    },
+    {
+      "epoch": 0.36,
+      "grad_norm": 0.035682205110788345,
+      "learning_rate": 0.00018454900998896996,
+      "loss": 0.6921,
+      "step": 653
+    },
+    {
+      "epoch": 0.36,
+      "grad_norm": 0.04010608792304993,
+      "learning_rate": 0.0001845021961073462,
+      "loss": 0.7686,
+      "step": 654
+    },
+    {
+      "epoch": 0.36,
+      "grad_norm": 0.04509714990854263,
+      "learning_rate": 0.00018445531737017298,
+      "loss": 0.7965,
+      "step": 655
+    },
+    {
+      "epoch": 0.36,
+      "grad_norm": 0.04406216740608215,
+      "learning_rate": 0.0001844083738134298,
+      "loss": 0.8243,
+      "step": 656
+    },
+    {
+      "epoch": 0.37,
+      "grad_norm": 0.03819778561592102,
+      "learning_rate": 0.00018436136547314593,
+      "loss": 0.6683,
+      "step": 657
+    },
+    {
+      "epoch": 0.37,
+      "grad_norm": 0.04046725109219551,
+      "learning_rate": 0.00018431429238540027,
+      "loss": 0.8624,
+      "step": 658
+    },
+    {
+      "epoch": 0.37,
+      "grad_norm": 0.04344424977898598,
+      "learning_rate": 0.00018426715458632153,
+      "loss": 0.7828,
+      "step": 659
+    },
+    {
+      "epoch": 0.37,
+      "grad_norm": 0.040390148758888245,
+      "learning_rate": 0.00018421995211208802,
+      "loss": 0.7757,
+      "step": 660
+    },
+    {
+      "epoch": 0.37,
+      "grad_norm": 0.03947126492857933,
+      "learning_rate": 0.00018417268499892767,
+      "loss": 0.6989,
+      "step": 661
+    },
+    {
+      "epoch": 0.37,
+      "grad_norm": 0.04135662689805031,
+      "learning_rate": 0.00018412535328311814,
+      "loss": 0.847,
+      "step": 662
+    },
+    {
+      "epoch": 0.37,
+      "grad_norm": 0.049522459506988525,
+      "learning_rate": 0.00018407795700098648,
+      "loss": 0.7971,
+      "step": 663
+    },
+    {
+      "epoch": 0.37,
+      "grad_norm": 0.040498074144124985,
+      "learning_rate": 0.0001840304961889095,
+      "loss": 0.7314,
+      "step": 664
+    },
+    {
+      "epoch": 0.37,
+      "grad_norm": 0.05009998753666878,
+      "learning_rate": 0.00018398297088331332,
+      "loss": 0.7435,
+      "step": 665
+    },
+    {
+      "epoch": 0.37,
+      "grad_norm": 0.03603378310799599,
+      "learning_rate": 0.0001839353811206738,
+      "loss": 0.7178,
+      "step": 666
+    },
+    {
+      "epoch": 0.37,
+      "grad_norm": 0.037989791482686996,
+      "learning_rate": 0.00018388772693751602,
+      "loss": 0.8039,
+      "step": 667
+    },
+    {
+      "epoch": 0.37,
+      "grad_norm": 0.03886855021119118,
+      "learning_rate": 0.00018384000837041477,
+      "loss": 0.7001,
+      "step": 668
+    },
+    {
+      "epoch": 0.37,
+      "grad_norm": 0.04233938828110695,
+      "learning_rate": 0.00018379222545599402,
+      "loss": 0.7929,
+      "step": 669
+    },
+    {
+      "epoch": 0.37,
+      "grad_norm": 0.04034002870321274,
+      "learning_rate": 0.00018374437823092724,
+      "loss": 0.6706,
+      "step": 670
+    },
+    {
+      "epoch": 0.37,
+      "grad_norm": 0.04359966889023781,
+      "learning_rate": 0.00018369646673193724,
+      "loss": 0.825,
+      "step": 671
+    },
+    {
+      "epoch": 0.37,
+      "grad_norm": 0.04203416034579277,
+      "learning_rate": 0.0001836484909957962,
+      "loss": 0.8887,
+      "step": 672
+    },
+    {
+      "epoch": 0.37,
+      "grad_norm": 0.04705547168850899,
+      "learning_rate": 0.0001836004510593255,
+      "loss": 0.7003,
+      "step": 673
+    },
+    {
+      "epoch": 0.37,
+      "grad_norm": 0.040125973522663116,
+      "learning_rate": 0.00018355234695939586,
+      "loss": 0.7106,
+      "step": 674
+    },
+    {
+      "epoch": 0.38,
+      "grad_norm": 0.040442850440740585,
+      "learning_rate": 0.0001835041787329273,
+      "loss": 0.6736,
+      "step": 675
+    },
+    {
+      "epoch": 0.38,
+      "grad_norm": 0.04361211508512497,
+      "learning_rate": 0.00018345594641688893,
+      "loss": 0.7427,
+      "step": 676
+    },
+    {
+      "epoch": 0.38,
+      "grad_norm": 0.039316337555646896,
+      "learning_rate": 0.0001834076500482992,
+      "loss": 0.7529,
+      "step": 677
+    },
+    {
+      "epoch": 0.38,
+      "grad_norm": 0.041209638118743896,
+      "learning_rate": 0.00018335928966422557,
+      "loss": 0.7345,
+      "step": 678
+    },
+    {
+      "epoch": 0.38,
+      "grad_norm": 0.04208279401063919,
+      "learning_rate": 0.0001833108653017847,
+      "loss": 0.7544,
+      "step": 679
+    },
+    {
+      "epoch": 0.38,
+      "grad_norm": 0.042629461735486984,
+      "learning_rate": 0.0001832623769981424,
+      "loss": 0.6676,
+      "step": 680
+    },
+    {
+      "epoch": 0.38,
+      "grad_norm": 0.038116682320833206,
+      "learning_rate": 0.00018321382479051347,
+      "loss": 0.705,
+      "step": 681
+    },
+    {
+      "epoch": 0.38,
+      "grad_norm": 0.03879912942647934,
+      "learning_rate": 0.0001831652087161618,
+      "loss": 0.7405,
+      "step": 682
+    },
+    {
+      "epoch": 0.38,
+      "grad_norm": 0.04380761831998825,
+      "learning_rate": 0.00018311652881240032,
+      "loss": 0.6791,
+      "step": 683
+    },
+    {
+      "epoch": 0.38,
+      "grad_norm": 0.03818592429161072,
+      "learning_rate": 0.00018306778511659085,
+      "loss": 0.6545,
+      "step": 684
+    },
+    {
+      "epoch": 0.38,
+      "grad_norm": 0.0418897345662117,
+      "learning_rate": 0.00018301897766614435,
+      "loss": 0.815,
+      "step": 685
+    },
+    {
+      "epoch": 0.38,
+      "grad_norm": 0.04104520007967949,
+      "learning_rate": 0.00018297010649852052,
+      "loss": 0.6821,
+      "step": 686
+    },
+    {
+      "epoch": 0.38,
+      "grad_norm": 0.04211374744772911,
+      "learning_rate": 0.0001829211716512281,
+      "loss": 0.778,
+      "step": 687
+    },
+    {
+      "epoch": 0.38,
+      "grad_norm": 0.045832112431526184,
+      "learning_rate": 0.00018287217316182458,
+      "loss": 0.8436,
+      "step": 688
+    },
+    {
+      "epoch": 0.38,
+      "grad_norm": 0.042089615017175674,
+      "learning_rate": 0.00018282311106791645,
+      "loss": 0.6899,
+      "step": 689
+    },
+    {
+      "epoch": 0.38,
+      "grad_norm": 0.039306432008743286,
+      "learning_rate": 0.0001827739854071589,
+      "loss": 0.6268,
+      "step": 690
+    },
+    {
+      "epoch": 0.38,
+      "grad_norm": 0.04297361150383949,
+      "learning_rate": 0.00018272479621725589,
+      "loss": 0.7461,
+      "step": 691
+    },
+    {
+      "epoch": 0.38,
+      "grad_norm": 0.04032345116138458,
+      "learning_rate": 0.00018267554353596025,
+      "loss": 0.797,
+      "step": 692
+    },
+    {
+      "epoch": 0.39,
+      "grad_norm": 0.04208484664559364,
+      "learning_rate": 0.0001826262274010735,
+      "loss": 0.8043,
+      "step": 693
+    },
+    {
+      "epoch": 0.39,
+      "grad_norm": 0.039476584643125534,
+      "learning_rate": 0.00018257684785044577,
+      "loss": 0.6953,
+      "step": 694
+    },
+    {
+      "epoch": 0.39,
+      "grad_norm": 0.040461137890815735,
+      "learning_rate": 0.00018252740492197598,
+      "loss": 0.754,
+      "step": 695
+    },
+    {
+      "epoch": 0.39,
+      "grad_norm": 0.04078752547502518,
+      "learning_rate": 0.00018247789865361166,
+      "loss": 0.6886,
+      "step": 696
+    },
+    {
+      "epoch": 0.39,
+      "grad_norm": 0.04459698870778084,
+      "learning_rate": 0.00018242832908334886,
+      "loss": 0.7355,
+      "step": 697
+    },
+    {
+      "epoch": 0.39,
+      "grad_norm": 0.039952464401721954,
+      "learning_rate": 0.00018237869624923236,
+      "loss": 0.7305,
+      "step": 698
+    },
+    {
+      "epoch": 0.39,
+      "grad_norm": 0.037101615220308304,
+      "learning_rate": 0.0001823290001893554,
+      "loss": 0.7426,
+      "step": 699
+    },
+    {
+      "epoch": 0.39,
+      "grad_norm": 0.050815049558877945,
+      "learning_rate": 0.0001822792409418598,
+      "loss": 0.6909,
+      "step": 700
+    },
+    {
+      "epoch": 0.39,
+      "grad_norm": 0.04514652490615845,
+      "learning_rate": 0.0001822294185449358,
+      "loss": 0.8794,
+      "step": 701
+    },
+    {
+      "epoch": 0.39,
+      "grad_norm": 0.04079204425215721,
+      "learning_rate": 0.00018217953303682223,
+      "loss": 0.6969,
+      "step": 702
+    },
+    {
+      "epoch": 0.39,
+      "grad_norm": 0.03969675302505493,
+      "learning_rate": 0.0001821295844558062,
+      "loss": 0.6827,
+      "step": 703
+    },
+    {
+      "epoch": 0.39,
+      "grad_norm": 0.04383305460214615,
+      "learning_rate": 0.0001820795728402234,
+      "loss": 0.8434,
+      "step": 704
+    },
+    {
+      "epoch": 0.39,
+      "grad_norm": 0.039204008877277374,
+      "learning_rate": 0.00018202949822845773,
+      "loss": 0.6717,
+      "step": 705
+    },
+    {
+      "epoch": 0.39,
+      "grad_norm": 0.03731399402022362,
+      "learning_rate": 0.00018197936065894157,
+      "loss": 0.5858,
+      "step": 706
+    },
+    {
+      "epoch": 0.39,
+      "grad_norm": 0.04733480513095856,
+      "learning_rate": 0.00018192916017015557,
+      "loss": 0.7383,
+      "step": 707
+    },
+    {
+      "epoch": 0.39,
+      "grad_norm": 0.04360615089535713,
+      "learning_rate": 0.00018187889680062864,
+      "loss": 0.716,
+      "step": 708
+    },
+    {
+      "epoch": 0.39,
+      "grad_norm": 0.040373023599386215,
+      "learning_rate": 0.00018182857058893803,
+      "loss": 0.6859,
+      "step": 709
+    },
+    {
+      "epoch": 0.39,
+      "grad_norm": 0.041170086711645126,
+      "learning_rate": 0.00018177818157370915,
+      "loss": 0.6335,
+      "step": 710
+    },
+    {
+      "epoch": 0.4,
+      "grad_norm": 0.039953090250492096,
+      "learning_rate": 0.00018172772979361567,
+      "loss": 0.6232,
+      "step": 711
+    },
+    {
+      "epoch": 0.4,
+      "grad_norm": 0.04438021779060364,
+      "learning_rate": 0.0001816772152873793,
+      "loss": 0.7427,
+      "step": 712
+    },
+    {
+      "epoch": 0.4,
+      "grad_norm": 0.037649158388376236,
+      "learning_rate": 0.00018162663809377012,
+      "loss": 0.6732,
+      "step": 713
+    },
+    {
+      "epoch": 0.4,
+      "grad_norm": 0.04502052068710327,
+      "learning_rate": 0.0001815759982516061,
+      "loss": 0.8635,
+      "step": 714
+    },
+    {
+      "epoch": 0.4,
+      "grad_norm": 0.04553660750389099,
+      "learning_rate": 0.00018152529579975343,
+      "loss": 0.7883,
+      "step": 715
+    },
+    {
+      "epoch": 0.4,
+      "grad_norm": 0.038996774703264236,
+      "learning_rate": 0.00018147453077712634,
+      "loss": 0.7818,
+      "step": 716
+    },
+    {
+      "epoch": 0.4,
+      "grad_norm": 0.03996030241250992,
+      "learning_rate": 0.000181423703222687,
+      "loss": 0.7777,
+      "step": 717
+    },
+    {
+      "epoch": 0.4,
+      "grad_norm": 0.04063892737030983,
+      "learning_rate": 0.0001813728131754456,
+      "loss": 0.7201,
+      "step": 718
+    },
+    {
+      "epoch": 0.4,
+      "grad_norm": 0.04045387730002403,
+      "learning_rate": 0.00018132186067446043,
+      "loss": 0.8018,
+      "step": 719
+    },
+    {
+      "epoch": 0.4,
+      "grad_norm": 0.042122699320316315,
+      "learning_rate": 0.0001812708457588375,
+      "loss": 0.7921,
+      "step": 720
+    },
+    {
+      "epoch": 0.4,
+      "grad_norm": 0.04201977327466011,
+      "learning_rate": 0.00018121976846773084,
+      "loss": 0.7049,
+      "step": 721
+    },
+    {
+      "epoch": 0.4,
+      "grad_norm": 0.03838672116398811,
+      "learning_rate": 0.00018116862884034243,
+      "loss": 0.6836,
+      "step": 722
+    },
+    {
+      "epoch": 0.4,
+      "grad_norm": 0.044829823076725006,
+      "learning_rate": 0.0001811174269159219,
+      "loss": 0.765,
+      "step": 723
+    },
+    {
+      "epoch": 0.4,
+      "grad_norm": 0.04038670286536217,
+      "learning_rate": 0.00018106616273376683,
+      "loss": 0.7528,
+      "step": 724
+    },
+    {
+      "epoch": 0.4,
+      "grad_norm": 0.041412707418203354,
+      "learning_rate": 0.00018101483633322255,
+      "loss": 0.8253,
+      "step": 725
+    },
+    {
+      "epoch": 0.4,
+      "grad_norm": 0.04251568764448166,
+      "learning_rate": 0.00018096344775368214,
+      "loss": 0.7483,
+      "step": 726
+    },
+    {
+      "epoch": 0.4,
+      "grad_norm": 0.043436773121356964,
+      "learning_rate": 0.0001809119970345864,
+      "loss": 0.7989,
+      "step": 727
+    },
+    {
+      "epoch": 0.4,
+      "grad_norm": 0.04096989333629608,
+      "learning_rate": 0.00018086048421542383,
+      "loss": 0.7947,
+      "step": 728
+    },
+    {
+      "epoch": 0.41,
+      "grad_norm": 0.04080723971128464,
+      "learning_rate": 0.0001808089093357306,
+      "loss": 0.7878,
+      "step": 729
+    },
+    {
+      "epoch": 0.41,
+      "grad_norm": 0.043036088347435,
+      "learning_rate": 0.0001807572724350905,
+      "loss": 0.7725,
+      "step": 730
+    },
+    {
+      "epoch": 0.41,
+      "grad_norm": 0.040440741926431656,
+      "learning_rate": 0.0001807055735531349,
+      "loss": 0.688,
+      "step": 731
+    },
+    {
+      "epoch": 0.41,
+      "grad_norm": 0.044196490198373795,
+      "learning_rate": 0.00018065381272954278,
+      "loss": 0.8786,
+      "step": 732
+    },
+    {
+      "epoch": 0.41,
+      "grad_norm": 0.042028769850730896,
+      "learning_rate": 0.00018060199000404062,
+      "loss": 0.8395,
+      "step": 733
+    },
+    {
+      "epoch": 0.41,
+      "grad_norm": 0.04810195043683052,
+      "learning_rate": 0.00018055010541640244,
+      "loss": 0.7793,
+      "step": 734
+    },
+    {
+      "epoch": 0.41,
+      "grad_norm": 0.03826632723212242,
+      "learning_rate": 0.0001804981590064498,
+      "loss": 0.6893,
+      "step": 735
+    },
+    {
+      "epoch": 0.41,
+      "grad_norm": 0.04325885698199272,
+      "learning_rate": 0.00018044615081405153,
+      "loss": 0.8006,
+      "step": 736
+    },
+    {
+      "epoch": 0.41,
+      "grad_norm": 0.039932068437337875,
+      "learning_rate": 0.00018039408087912402,
+      "loss": 0.6664,
+      "step": 737
+    },
+    {
+      "epoch": 0.41,
+      "grad_norm": 0.04278779402375221,
+      "learning_rate": 0.00018034194924163103,
+      "loss": 0.7397,
+      "step": 738
+    },
+    {
+      "epoch": 0.41,
+      "grad_norm": 0.03901227191090584,
+      "learning_rate": 0.0001802897559415837,
+      "loss": 0.5662,
+      "step": 739
+    },
+    {
+      "epoch": 0.41,
+      "grad_norm": 0.043960485607385635,
+      "learning_rate": 0.0001802375010190404,
+      "loss": 0.7809,
+      "step": 740
+    },
+    {
+      "epoch": 0.41,
+      "grad_norm": 0.04186437651515007,
+      "learning_rate": 0.0001801851845141069,
+      "loss": 0.6976,
+      "step": 741
+    },
+    {
+      "epoch": 0.41,
+      "grad_norm": 0.041978247463703156,
+      "learning_rate": 0.00018013280646693612,
+      "loss": 0.7146,
+      "step": 742
+    },
+    {
+      "epoch": 0.41,
+      "grad_norm": 0.04281838610768318,
+      "learning_rate": 0.00018008036691772835,
+      "loss": 0.7627,
+      "step": 743
+    },
+    {
+      "epoch": 0.41,
+      "grad_norm": 0.0473211370408535,
+      "learning_rate": 0.00018002786590673098,
+      "loss": 0.8042,
+      "step": 744
+    },
+    {
+      "epoch": 0.41,
+      "grad_norm": 0.04652298986911774,
+      "learning_rate": 0.0001799753034742386,
+      "loss": 0.731,
+      "step": 745
+    },
+    {
+      "epoch": 0.41,
+      "grad_norm": 0.039468780159950256,
+      "learning_rate": 0.000179922679660593,
+      "loss": 0.7135,
+      "step": 746
+    },
+    {
+      "epoch": 0.42,
+      "grad_norm": 0.04733370244503021,
+      "learning_rate": 0.00017986999450618295,
+      "loss": 0.8023,
+      "step": 747
+    },
+    {
+      "epoch": 0.42,
+      "grad_norm": 0.04231337085366249,
+      "learning_rate": 0.00017981724805144443,
+      "loss": 0.7528,
+      "step": 748
+    },
+    {
+      "epoch": 0.42,
+      "grad_norm": 0.04062468931078911,
+      "learning_rate": 0.00017976444033686043,
+      "loss": 0.7331,
+      "step": 749
+    },
+    {
+      "epoch": 0.42,
+      "grad_norm": 0.04496033489704132,
+      "learning_rate": 0.0001797115714029609,
+      "loss": 0.6968,
+      "step": 750
+    },
+    {
+      "epoch": 0.42,
+      "grad_norm": 0.04302246496081352,
+      "learning_rate": 0.00017965864129032284,
+      "loss": 0.8159,
+      "step": 751
+    },
+    {
+      "epoch": 0.42,
+      "grad_norm": 0.04262755811214447,
+      "learning_rate": 0.00017960565003957018,
+      "loss": 0.7148,
+      "step": 752
+    },
+    {
+      "epoch": 0.42,
+      "grad_norm": 0.040617603808641434,
+      "learning_rate": 0.00017955259769137375,
+      "loss": 0.7302,
+      "step": 753
+    },
+    {
+      "epoch": 0.42,
+      "grad_norm": 0.04194014146924019,
+      "learning_rate": 0.00017949948428645134,
+      "loss": 0.6785,
+      "step": 754
+    },
+    {
+      "epoch": 0.42,
+      "grad_norm": 0.04514269158244133,
+      "learning_rate": 0.00017944630986556752,
+      "loss": 0.748,
+      "step": 755
+    },
+    {
+      "epoch": 0.42,
+      "grad_norm": 0.042905841022729874,
+      "learning_rate": 0.0001793930744695338,
+      "loss": 0.7357,
+      "step": 756
+    },
+    {
+      "epoch": 0.42,
+      "grad_norm": 0.04359228163957596,
+      "learning_rate": 0.00017933977813920833,
+      "loss": 0.7426,
+      "step": 757
+    },
+    {
+      "epoch": 0.42,
+      "grad_norm": 0.040619876235723495,
+      "learning_rate": 0.00017928642091549613,
+      "loss": 0.7278,
+      "step": 758
+    },
+    {
+      "epoch": 0.42,
+      "grad_norm": 0.039674315601587296,
+      "learning_rate": 0.000179233002839349,
+      "loss": 0.7459,
+      "step": 759
+    },
+    {
+      "epoch": 0.42,
+      "grad_norm": 0.039900798350572586,
+      "learning_rate": 0.00017917952395176537,
+      "loss": 0.6963,
+      "step": 760
+    },
+    {
+      "epoch": 0.42,
+      "grad_norm": 0.039163414388895035,
+      "learning_rate": 0.0001791259842937903,
+      "loss": 0.652,
+      "step": 761
+    },
+    {
+      "epoch": 0.42,
+      "grad_norm": 0.04467619210481644,
+      "learning_rate": 0.0001790723839065156,
+      "loss": 0.585,
+      "step": 762
+    },
+    {
+      "epoch": 0.42,
+      "grad_norm": 0.03953075036406517,
+      "learning_rate": 0.00017901872283107966,
+      "loss": 0.7537,
+      "step": 763
+    },
+    {
+      "epoch": 0.42,
+      "grad_norm": 0.0408446304500103,
+      "learning_rate": 0.0001789650011086674,
+      "loss": 0.7833,
+      "step": 764
+    },
+    {
+      "epoch": 0.43,
+      "grad_norm": 0.05530129000544548,
+      "learning_rate": 0.00017891121878051025,
+      "loss": 0.8796,
+      "step": 765
+    },
+    {
+      "epoch": 0.43,
+      "grad_norm": 0.03987415134906769,
+      "learning_rate": 0.00017885737588788634,
+      "loss": 0.6762,
+      "step": 766
+    },
+    {
+      "epoch": 0.43,
+      "grad_norm": 0.042246054857969284,
+      "learning_rate": 0.00017880347247212005,
+      "loss": 0.7387,
+      "step": 767
+    },
+    {
+      "epoch": 0.43,
+      "grad_norm": 0.04334084317088127,
+      "learning_rate": 0.00017874950857458242,
+      "loss": 0.7526,
+      "step": 768
+    },
+    {
+      "epoch": 0.43,
+      "grad_norm": 0.04815899208188057,
+      "learning_rate": 0.00017869548423669077,
+      "loss": 0.8278,
+      "step": 769
+    },
+    {
+      "epoch": 0.43,
+      "grad_norm": 0.05985521152615547,
+      "learning_rate": 0.00017864139949990885,
+      "loss": 0.6726,
+      "step": 770
+    },
+    {
+      "epoch": 0.43,
+      "grad_norm": 0.038774825632572174,
+      "learning_rate": 0.00017858725440574676,
+      "loss": 0.6549,
+      "step": 771
+    },
+    {
+      "epoch": 0.43,
+      "grad_norm": 0.09850037842988968,
+      "learning_rate": 0.00017853304899576093,
+      "loss": 1.1263,
+      "step": 772
+    },
+    {
+      "epoch": 0.43,
+      "grad_norm": 0.0418662466108799,
+      "learning_rate": 0.0001784787833115541,
+      "loss": 0.6866,
+      "step": 773
+    },
+    {
+      "epoch": 0.43,
+      "grad_norm": 0.04924052208662033,
+      "learning_rate": 0.00017842445739477532,
+      "loss": 0.7591,
+      "step": 774
+    },
+    {
+      "epoch": 0.43,
+      "grad_norm": 0.04036710783839226,
+      "learning_rate": 0.0001783700712871197,
+      "loss": 0.7177,
+      "step": 775
+    },
+    {
+      "epoch": 0.43,
+      "grad_norm": 0.04530615732073784,
+      "learning_rate": 0.00017831562503032867,
+      "loss": 0.8464,
+      "step": 776
+    },
+    {
+      "epoch": 0.43,
+      "grad_norm": 0.04415806382894516,
+      "learning_rate": 0.00017826111866618986,
+      "loss": 0.6993,
+      "step": 777
+    },
+    {
+      "epoch": 0.43,
+      "grad_norm": 0.04042017087340355,
+      "learning_rate": 0.00017820655223653692,
+      "loss": 0.6305,
+      "step": 778
+    },
+    {
+      "epoch": 0.43,
+      "grad_norm": 0.04573072865605354,
+      "learning_rate": 0.0001781519257832497,
+      "loss": 0.8541,
+      "step": 779
+    },
+    {
+      "epoch": 0.43,
+      "grad_norm": 0.04222429171204567,
+      "learning_rate": 0.00017809723934825405,
+      "loss": 0.6579,
+      "step": 780
+    },
+    {
+      "epoch": 0.43,
+      "grad_norm": 0.04141012579202652,
+      "learning_rate": 0.0001780424929735219,
+      "loss": 0.7556,
+      "step": 781
+    },
+    {
+      "epoch": 0.43,
+      "grad_norm": 0.049307554960250854,
+      "learning_rate": 0.00017798768670107114,
+      "loss": 0.6392,
+      "step": 782
+    },
+    {
+      "epoch": 0.44,
+      "grad_norm": 0.0450025275349617,
+      "learning_rate": 0.0001779328205729657,
+      "loss": 0.7718,
+      "step": 783
+    },
+    {
+      "epoch": 0.44,
+      "grad_norm": 0.043787937611341476,
+      "learning_rate": 0.00017787789463131538,
+      "loss": 0.7086,
+      "step": 784
+    },
+    {
+      "epoch": 0.44,
+      "grad_norm": 0.03968477621674538,
+      "learning_rate": 0.0001778229089182759,
+      "loss": 0.6744,
+      "step": 785
+    },
+    {
+      "epoch": 0.44,
+      "grad_norm": 0.044153615832328796,
+      "learning_rate": 0.00017776786347604892,
+      "loss": 0.7169,
+      "step": 786
+    },
+    {
+      "epoch": 0.44,
+      "grad_norm": 0.04364906996488571,
+      "learning_rate": 0.00017771275834688183,
+      "loss": 0.7977,
+      "step": 787
+    },
+    {
+      "epoch": 0.44,
+      "grad_norm": 0.04209036007523537,
+      "learning_rate": 0.00017765759357306796,
+      "loss": 0.6939,
+      "step": 788
+    },
+    {
+      "epoch": 0.44,
+      "grad_norm": 0.04396096616983414,
+      "learning_rate": 0.00017760236919694627,
+      "loss": 0.7431,
+      "step": 789
+    },
+    {
+      "epoch": 0.44,
+      "grad_norm": 0.039003144949674606,
+      "learning_rate": 0.00017754708526090157,
+      "loss": 0.6491,
+      "step": 790
+    },
+    {
+      "epoch": 0.44,
+      "grad_norm": 0.04044824466109276,
+      "learning_rate": 0.00017749174180736442,
+      "loss": 0.6875,
+      "step": 791
+    },
+    {
+      "epoch": 0.44,
+      "grad_norm": 0.04009506106376648,
+      "learning_rate": 0.0001774363388788109,
+      "loss": 0.6727,
+      "step": 792
+    },
+    {
+      "epoch": 0.44,
+      "grad_norm": 0.044321801513433456,
+      "learning_rate": 0.00017738087651776287,
+      "loss": 0.6705,
+      "step": 793
+    },
+    {
+      "epoch": 0.44,
+      "grad_norm": 0.047294605523347855,
+      "learning_rate": 0.00017732535476678777,
+      "loss": 0.8299,
+      "step": 794
+    },
+    {
+      "epoch": 0.44,
+      "grad_norm": 0.04393976926803589,
+      "learning_rate": 0.00017726977366849865,
+      "loss": 0.8299,
+      "step": 795
+    },
+    {
+      "epoch": 0.44,
+      "grad_norm": 0.038830842822790146,
+      "learning_rate": 0.000177214133265554,
+      "loss": 0.768,
+      "step": 796
+    },
+    {
+      "epoch": 0.44,
+      "grad_norm": 0.04319118335843086,
+      "learning_rate": 0.00017715843360065797,
+      "loss": 0.83,
+      "step": 797
+    },
+    {
+      "epoch": 0.44,
+      "grad_norm": 0.04209451377391815,
+      "learning_rate": 0.00017710267471656016,
+      "loss": 0.6847,
+      "step": 798
+    },
+    {
+      "epoch": 0.44,
+      "grad_norm": 0.04092387482523918,
+      "learning_rate": 0.00017704685665605546,
+      "loss": 0.7254,
+      "step": 799
+    },
+    {
+      "epoch": 0.44,
+      "grad_norm": 0.041155993938446045,
+      "learning_rate": 0.00017699097946198444,
+      "loss": 0.6125,
+      "step": 800
+    },
+    {
+      "epoch": 0.45,
+      "grad_norm": 0.03681296482682228,
+      "learning_rate": 0.00017693504317723284,
+      "loss": 0.6512,
+      "step": 801
+    },
+    {
+      "epoch": 0.45,
+      "grad_norm": 0.04595888406038284,
+      "learning_rate": 0.00017687904784473188,
+      "loss": 0.7719,
+      "step": 802
+    },
+    {
+      "epoch": 0.45,
+      "grad_norm": 0.04086902737617493,
+      "learning_rate": 0.00017682299350745803,
+      "loss": 0.6339,
+      "step": 803
+    },
+    {
+      "epoch": 0.45,
+      "grad_norm": 0.04581453278660774,
+      "learning_rate": 0.00017676688020843307,
+      "loss": 0.6704,
+      "step": 804
+    },
+    {
+      "epoch": 0.45,
+      "grad_norm": 0.041530586779117584,
+      "learning_rate": 0.00017671070799072402,
+      "loss": 0.79,
+      "step": 805
+    },
+    {
+      "epoch": 0.45,
+      "grad_norm": 0.04776669666171074,
+      "learning_rate": 0.0001766544768974432,
+      "loss": 0.782,
+      "step": 806
+    },
+    {
+      "epoch": 0.45,
+      "grad_norm": 0.041267361491918564,
+      "learning_rate": 0.00017659818697174796,
+      "loss": 0.6647,
+      "step": 807
+    },
+    {
+      "epoch": 0.45,
+      "grad_norm": 0.04318636655807495,
+      "learning_rate": 0.00017654183825684092,
+      "loss": 0.7395,
+      "step": 808
+    },
+    {
+      "epoch": 0.45,
+      "grad_norm": 0.04291536659002304,
+      "learning_rate": 0.00017648543079596982,
+      "loss": 0.7857,
+      "step": 809
+    },
+    {
+      "epoch": 0.45,
+      "grad_norm": 0.04914848133921623,
+      "learning_rate": 0.00017642896463242747,
+      "loss": 0.8949,
+      "step": 810
+    },
+    {
+      "epoch": 0.45,
+      "grad_norm": 0.042548660188913345,
+      "learning_rate": 0.00017637243980955168,
+      "loss": 0.6978,
+      "step": 811
+    },
+    {
+      "epoch": 0.45,
+      "grad_norm": 0.044454239308834076,
+      "learning_rate": 0.00017631585637072536,
+      "loss": 0.7645,
+      "step": 812
+    },
+    {
+      "epoch": 0.45,
+      "grad_norm": 0.044261105358600616,
+      "learning_rate": 0.00017625921435937637,
+      "loss": 0.704,
+      "step": 813
+    },
+    {
+      "epoch": 0.45,
+      "grad_norm": 0.03984127193689346,
+      "learning_rate": 0.00017620251381897752,
+      "loss": 0.6803,
+      "step": 814
+    },
+    {
+      "epoch": 0.45,
+      "grad_norm": 0.04928235337138176,
+      "learning_rate": 0.00017614575479304662,
+      "loss": 0.6964,
+      "step": 815
+    },
+    {
+      "epoch": 0.45,
+      "grad_norm": 0.03964222967624664,
+      "learning_rate": 0.00017608893732514616,
+      "loss": 0.7508,
+      "step": 816
+    },
+    {
+      "epoch": 0.45,
+      "grad_norm": 0.04107435792684555,
+      "learning_rate": 0.00017603206145888373,
+      "loss": 0.6977,
+      "step": 817
+    },
+    {
+      "epoch": 0.45,
+      "grad_norm": 0.03979944810271263,
+      "learning_rate": 0.00017597512723791162,
+      "loss": 0.6823,
+      "step": 818
+    },
+    {
+      "epoch": 0.46,
+      "grad_norm": 0.04214751720428467,
+      "learning_rate": 0.00017591813470592692,
+      "loss": 0.756,
+      "step": 819
+    },
+    {
+      "epoch": 0.46,
+      "grad_norm": 0.04152442887425423,
+      "learning_rate": 0.00017586108390667144,
+      "loss": 0.6977,
+      "step": 820
+    },
+    {
+      "epoch": 0.46,
+      "grad_norm": 0.04409961402416229,
+      "learning_rate": 0.00017580397488393176,
+      "loss": 0.7103,
+      "step": 821
+    },
+    {
+      "epoch": 0.46,
+      "grad_norm": 0.04341358318924904,
+      "learning_rate": 0.00017574680768153917,
+      "loss": 0.7553,
+      "step": 822
+    },
+    {
+      "epoch": 0.46,
+      "grad_norm": 0.040573328733444214,
+      "learning_rate": 0.00017568958234336952,
+      "loss": 0.6273,
+      "step": 823
+    },
+    {
+      "epoch": 0.46,
+      "grad_norm": 0.04173550009727478,
+      "learning_rate": 0.00017563229891334338,
+      "loss": 0.6156,
+      "step": 824
+    },
+    {
+      "epoch": 0.46,
+      "grad_norm": 0.043724823743104935,
+      "learning_rate": 0.00017557495743542585,
+      "loss": 0.681,
+      "step": 825
+    },
+    {
+      "epoch": 0.46,
+      "grad_norm": 0.04628165438771248,
+      "learning_rate": 0.00017551755795362656,
+      "loss": 0.707,
+      "step": 826
+    },
+    {
+      "epoch": 0.46,
+      "grad_norm": 0.0463298037648201,
+      "learning_rate": 0.00017546010051199971,
+      "loss": 0.7532,
+      "step": 827
+    },
+    {
+      "epoch": 0.46,
+      "grad_norm": 0.04120299220085144,
+      "learning_rate": 0.00017540258515464396,
+      "loss": 0.6755,
+      "step": 828
+    },
+    {
+      "epoch": 0.46,
+      "grad_norm": 0.04657519608736038,
+      "learning_rate": 0.00017534501192570246,
+      "loss": 0.794,
+      "step": 829
+    },
+    {
+      "epoch": 0.46,
+      "grad_norm": 0.04502912238240242,
+      "learning_rate": 0.0001752873808693627,
+      "loss": 0.7433,
+      "step": 830
+    },
+    {
+      "epoch": 0.46,
+      "grad_norm": 0.05056634172797203,
+      "learning_rate": 0.00017522969202985657,
+      "loss": 0.8462,
+      "step": 831
+    },
+    {
+      "epoch": 0.46,
+      "grad_norm": 0.04288472607731819,
+      "learning_rate": 0.00017517194545146037,
+      "loss": 0.6889,
+      "step": 832
+    },
+    {
+      "epoch": 0.46,
+      "grad_norm": 0.044761136174201965,
+      "learning_rate": 0.0001751141411784947,
+      "loss": 0.6657,
+      "step": 833
+    },
+    {
+      "epoch": 0.46,
+      "grad_norm": 0.04813416302204132,
+      "learning_rate": 0.00017505627925532442,
+      "loss": 0.7993,
+      "step": 834
+    },
+    {
+      "epoch": 0.46,
+      "grad_norm": 0.04681932181119919,
+      "learning_rate": 0.00017499835972635856,
+      "loss": 0.8888,
+      "step": 835
+    },
+    {
+      "epoch": 0.46,
+      "grad_norm": 0.04293239116668701,
+      "learning_rate": 0.0001749403826360505,
+      "loss": 0.7787,
+      "step": 836
+    },
+    {
+      "epoch": 0.47,
+      "grad_norm": 0.04279159754514694,
+      "learning_rate": 0.0001748823480288977,
+      "loss": 0.6838,
+      "step": 837
+    },
+    {
+      "epoch": 0.47,
+      "grad_norm": 0.04303467273712158,
+      "learning_rate": 0.00017482425594944184,
+      "loss": 0.7993,
+      "step": 838
+    },
+    {
+      "epoch": 0.47,
+      "grad_norm": 0.044777967035770416,
+      "learning_rate": 0.0001747661064422686,
+      "loss": 0.8355,
+      "step": 839
+    },
+    {
+      "epoch": 0.47,
+      "grad_norm": 0.04495497792959213,
+      "learning_rate": 0.00017470789955200788,
+      "loss": 0.8078,
+      "step": 840
+    },
+    {
+      "epoch": 0.47,
+      "grad_norm": 0.05037945136427879,
+      "learning_rate": 0.00017464963532333352,
+      "loss": 0.6933,
+      "step": 841
+    },
+    {
+      "epoch": 0.47,
+      "grad_norm": 0.04452716186642647,
+      "learning_rate": 0.0001745913138009634,
+      "loss": 0.7993,
+      "step": 842
+    },
+    {
+      "epoch": 0.47,
+      "grad_norm": 0.042573876678943634,
+      "learning_rate": 0.0001745329350296593,
+      "loss": 0.6218,
+      "step": 843
+    },
+    {
+      "epoch": 0.47,
+      "grad_norm": 0.04359611123800278,
+      "learning_rate": 0.00017447449905422713,
+      "loss": 0.7138,
+      "step": 844
+    },
+    {
+      "epoch": 0.47,
+      "grad_norm": 0.03893902152776718,
+      "learning_rate": 0.00017441600591951647,
+      "loss": 0.5625,
+      "step": 845
+    },
+    {
+      "epoch": 0.47,
+      "grad_norm": 0.04326329007744789,
+      "learning_rate": 0.00017435745567042095,
+      "loss": 0.808,
+      "step": 846
+    },
+    {
+      "epoch": 0.47,
+      "grad_norm": 0.04221475496888161,
+      "learning_rate": 0.0001742988483518779,
+      "loss": 0.6852,
+      "step": 847
+    },
+    {
+      "epoch": 0.47,
+      "grad_norm": 0.04148104414343834,
+      "learning_rate": 0.0001742401840088686,
+      "loss": 0.7226,
+      "step": 848
+    },
+    {
+      "epoch": 0.47,
+      "grad_norm": 0.043071821331977844,
+      "learning_rate": 0.00017418146268641794,
+      "loss": 0.7685,
+      "step": 849
+    },
+    {
+      "epoch": 0.47,
+      "grad_norm": 0.04110527038574219,
+      "learning_rate": 0.00017412268442959465,
+      "loss": 0.6696,
+      "step": 850
+    },
+    {
+      "epoch": 0.47,
+      "grad_norm": 0.04123936966061592,
+      "learning_rate": 0.00017406384928351113,
+      "loss": 0.6247,
+      "step": 851
+    },
+    {
+      "epoch": 0.47,
+      "grad_norm": 0.04614235460758209,
+      "learning_rate": 0.00017400495729332337,
+      "loss": 0.7239,
+      "step": 852
+    },
+    {
+      "epoch": 0.47,
+      "grad_norm": 0.04393448308110237,
+      "learning_rate": 0.00017394600850423114,
+      "loss": 0.6983,
+      "step": 853
+    },
+    {
+      "epoch": 0.47,
+      "grad_norm": 0.04823063313961029,
+      "learning_rate": 0.00017388700296147765,
+      "loss": 0.8802,
+      "step": 854
+    },
+    {
+      "epoch": 0.48,
+      "grad_norm": 0.04590068385004997,
+      "learning_rate": 0.00017382794071034975,
+      "loss": 0.764,
+      "step": 855
+    },
+    {
+      "epoch": 0.48,
+      "grad_norm": 0.043450817465782166,
+      "learning_rate": 0.00017376882179617783,
+      "loss": 0.6518,
+      "step": 856
+    },
+    {
+      "epoch": 0.48,
+      "grad_norm": 0.0406193882226944,
+      "learning_rate": 0.00017370964626433567,
+      "loss": 0.6845,
+      "step": 857
+    },
+    {
+      "epoch": 0.48,
+      "grad_norm": 0.0440056286752224,
+      "learning_rate": 0.00017365041416024065,
+      "loss": 0.7249,
+      "step": 858
+    },
+    {
+      "epoch": 0.48,
+      "grad_norm": 0.044533830136060715,
+      "learning_rate": 0.00017359112552935347,
+      "loss": 0.7045,
+      "step": 859
+    },
+    {
+      "epoch": 0.48,
+      "grad_norm": 0.04595007747411728,
+      "learning_rate": 0.00017353178041717814,
+      "loss": 0.7224,
+      "step": 860
+    },
+    {
+      "epoch": 0.48,
+      "grad_norm": 0.04445382580161095,
+      "learning_rate": 0.00017347237886926225,
+      "loss": 0.7391,
+      "step": 861
+    },
+    {
+      "epoch": 0.48,
+      "grad_norm": 0.044705115258693695,
+      "learning_rate": 0.0001734129209311965,
+      "loss": 0.8049,
+      "step": 862
+    },
+    {
+      "epoch": 0.48,
+      "grad_norm": 0.03983869031071663,
+      "learning_rate": 0.00017335340664861493,
+      "loss": 0.6245,
+      "step": 863
+    },
+    {
+      "epoch": 0.48,
+      "grad_norm": 0.03903631865978241,
+      "learning_rate": 0.00017329383606719481,
+      "loss": 0.6075,
+      "step": 864
+    },
+    {
+      "epoch": 0.48,
+      "grad_norm": 0.0470118448138237,
+      "learning_rate": 0.00017323420923265673,
+      "loss": 0.7978,
+      "step": 865
+    },
+    {
+      "epoch": 0.48,
+      "grad_norm": 0.046046625822782516,
+      "learning_rate": 0.00017317452619076428,
+      "loss": 0.7536,
+      "step": 866
+    },
+    {
+      "epoch": 0.48,
+      "grad_norm": 0.044642165303230286,
+      "learning_rate": 0.00017311478698732433,
+      "loss": 0.6868,
+      "step": 867
+    },
+    {
+      "epoch": 0.48,
+      "grad_norm": 0.043679915368556976,
+      "learning_rate": 0.0001730549916681868,
+      "loss": 0.6715,
+      "step": 868
+    },
+    {
+      "epoch": 0.48,
+      "grad_norm": 0.04374048486351967,
+      "learning_rate": 0.0001729951402792446,
+      "loss": 0.7951,
+      "step": 869
+    },
+    {
+      "epoch": 0.48,
+      "grad_norm": 0.04199199378490448,
+      "learning_rate": 0.00017293523286643386,
+      "loss": 0.6943,
+      "step": 870
+    },
+    {
+      "epoch": 0.48,
+      "grad_norm": 0.045180678367614746,
+      "learning_rate": 0.00017287526947573354,
+      "loss": 0.7037,
+      "step": 871
+    },
+    {
+      "epoch": 0.48,
+      "grad_norm": 0.04861563816666603,
+      "learning_rate": 0.0001728152501531656,
+      "loss": 0.808,
+      "step": 872
+    },
+    {
+      "epoch": 0.49,
+      "grad_norm": 0.06491879373788834,
+      "learning_rate": 0.000172755174944795,
+      "loss": 0.7232,
+      "step": 873
+    },
+    {
+      "epoch": 0.49,
+      "grad_norm": 0.056284647434949875,
+      "learning_rate": 0.0001726950438967295,
+      "loss": 0.7938,
+      "step": 874
+    },
+    {
+      "epoch": 0.49,
+      "grad_norm": 0.04053090140223503,
+      "learning_rate": 0.00017263485705511984,
+      "loss": 0.67,
+      "step": 875
+    },
+    {
+      "epoch": 0.49,
+      "grad_norm": 0.04802017658948898,
+      "learning_rate": 0.00017257461446615942,
+      "loss": 0.8086,
+      "step": 876
+    },
+    {
+      "epoch": 0.49,
+      "grad_norm": 0.055473774671554565,
+      "learning_rate": 0.00017251431617608452,
+      "loss": 0.6176,
+      "step": 877
+    },
+    {
+      "epoch": 0.49,
+      "grad_norm": 0.04549332708120346,
+      "learning_rate": 0.0001724539622311742,
+      "loss": 0.6832,
+      "step": 878
+    },
+    {
+      "epoch": 0.49,
+      "grad_norm": 0.040823038667440414,
+      "learning_rate": 0.00017239355267775018,
+      "loss": 0.6617,
+      "step": 879
+    },
+    {
+      "epoch": 0.49,
+      "grad_norm": 0.045911453664302826,
+      "learning_rate": 0.00017233308756217682,
+      "loss": 0.7098,
+      "step": 880
+    },
+    {
+      "epoch": 0.49,
+      "grad_norm": 0.044026948511600494,
+      "learning_rate": 0.00017227256693086123,
+      "loss": 0.7281,
+      "step": 881
+    },
+    {
+      "epoch": 0.49,
+      "grad_norm": 0.04080135002732277,
+      "learning_rate": 0.00017221199083025307,
+      "loss": 0.6427,
+      "step": 882
+    },
+    {
+      "epoch": 0.49,
+      "grad_norm": 0.04171103984117508,
+      "learning_rate": 0.0001721513593068446,
+      "loss": 0.77,
+      "step": 883
+    },
+    {
+      "epoch": 0.49,
+      "grad_norm": 0.04672902449965477,
+      "learning_rate": 0.00017209067240717057,
+      "loss": 0.6622,
+      "step": 884
+    },
+    {
+      "epoch": 0.49,
+      "grad_norm": 0.047331150621175766,
+      "learning_rate": 0.00017202993017780823,
+      "loss": 0.6539,
+      "step": 885
+    },
+    {
+      "epoch": 0.49,
+      "grad_norm": 0.041491780430078506,
+      "learning_rate": 0.0001719691326653774,
+      "loss": 0.693,
+      "step": 886
+    },
+    {
+      "epoch": 0.49,
+      "grad_norm": 0.04602917283773422,
+      "learning_rate": 0.00017190827991654014,
+      "loss": 0.6925,
+      "step": 887
+    },
+    {
+      "epoch": 0.49,
+      "grad_norm": 0.04347170516848564,
+      "learning_rate": 0.00017184737197800115,
+      "loss": 0.6765,
+      "step": 888
+    },
+    {
+      "epoch": 0.49,
+      "grad_norm": 0.04298747330904007,
+      "learning_rate": 0.00017178640889650723,
+      "loss": 0.6898,
+      "step": 889
+    },
+    {
+      "epoch": 0.49,
+      "grad_norm": 0.04869624972343445,
+      "learning_rate": 0.0001717253907188477,
+      "loss": 0.8587,
+      "step": 890
+    },
+    {
+      "epoch": 0.5,
+      "grad_norm": 0.04220546409487724,
+      "learning_rate": 0.0001716643174918541,
+      "loss": 0.6139,
+      "step": 891
+    },
+    {
+      "epoch": 0.5,
+      "grad_norm": 0.04683152586221695,
+      "learning_rate": 0.00017160318926240015,
+      "loss": 0.7845,
+      "step": 892
+    },
+    {
+      "epoch": 0.5,
+      "grad_norm": 0.04245654493570328,
+      "learning_rate": 0.0001715420060774019,
+      "loss": 0.6194,
+      "step": 893
+    },
+    {
+      "epoch": 0.5,
+      "grad_norm": 0.051294971257448196,
+      "learning_rate": 0.00017148076798381755,
+      "loss": 0.7762,
+      "step": 894
+    },
+    {
+      "epoch": 0.5,
+      "grad_norm": 0.05587690323591232,
+      "learning_rate": 0.00017141947502864738,
+      "loss": 0.7732,
+      "step": 895
+    },
+    {
+      "epoch": 0.5,
+      "grad_norm": 0.04731575399637222,
+      "learning_rate": 0.00017135812725893381,
+      "loss": 0.7229,
+      "step": 896
+    },
+    {
+      "epoch": 0.5,
+      "grad_norm": 0.04307183623313904,
+      "learning_rate": 0.00017129672472176134,
+      "loss": 0.6533,
+      "step": 897
+    },
+    {
+      "epoch": 0.5,
+      "grad_norm": 0.044035330414772034,
+      "learning_rate": 0.00017123526746425652,
+      "loss": 0.7754,
+      "step": 898
+    },
+    {
+      "epoch": 0.5,
+      "grad_norm": 0.04161735251545906,
+      "learning_rate": 0.00017117375553358786,
+      "loss": 0.6989,
+      "step": 899
+    },
+    {
+      "epoch": 0.5,
+      "grad_norm": 0.0438537523150444,
+      "learning_rate": 0.00017111218897696587,
+      "loss": 0.742,
+      "step": 900
+    },
+    {
+      "epoch": 0.5,
+      "grad_norm": 0.04731033742427826,
+      "learning_rate": 0.00017105056784164294,
+      "loss": 0.6689,
+      "step": 901
+    },
+    {
+      "epoch": 0.5,
+      "grad_norm": 0.04555543512105942,
+      "learning_rate": 0.00017098889217491338,
+      "loss": 0.7309,
+      "step": 902
+    },
+    {
+      "epoch": 0.5,
+      "grad_norm": 0.041630081832408905,
+      "learning_rate": 0.00017092716202411336,
+      "loss": 0.6845,
+      "step": 903
+    },
+    {
+      "epoch": 0.5,
+      "grad_norm": 0.045560967177152634,
+      "learning_rate": 0.00017086537743662086,
+      "loss": 0.6913,
+      "step": 904
+    },
+    {
+      "epoch": 0.5,
+      "grad_norm": 0.03997902199625969,
+      "learning_rate": 0.00017080353845985559,
+      "loss": 0.6016,
+      "step": 905
+    },
+    {
+      "epoch": 0.5,
+      "grad_norm": 0.04500821977853775,
+      "learning_rate": 0.0001707416451412791,
+      "loss": 0.7355,
+      "step": 906
+    },
+    {
+      "epoch": 0.5,
+      "grad_norm": 0.04442552849650383,
+      "learning_rate": 0.00017067969752839458,
+      "loss": 0.6525,
+      "step": 907
+    },
+    {
+      "epoch": 0.51,
+      "grad_norm": 0.04527292773127556,
+      "learning_rate": 0.0001706176956687469,
+      "loss": 0.6991,
+      "step": 908
+    },
+    {
+      "epoch": 0.51,
+      "grad_norm": 0.04470342397689819,
+      "learning_rate": 0.00017055563960992256,
+      "loss": 0.7352,
+      "step": 909
+    },
+    {
+      "epoch": 0.51,
+      "grad_norm": 0.04706728830933571,
+      "learning_rate": 0.00017049352939954967,
+      "loss": 0.7081,
+      "step": 910
+    },
+    {
+      "epoch": 0.51,
+      "grad_norm": 0.049026861786842346,
+      "learning_rate": 0.00017043136508529793,
+      "loss": 0.833,
+      "step": 911
+    },
+    {
+      "epoch": 0.51,
+      "grad_norm": 0.04529300704598427,
+      "learning_rate": 0.00017036914671487852,
+      "loss": 0.6624,
+      "step": 912
+    },
+    {
+      "epoch": 0.51,
+      "grad_norm": 0.05243083834648132,
+      "learning_rate": 0.0001703068743360441,
+      "loss": 0.9235,
+      "step": 913
+    },
+    {
+      "epoch": 0.51,
+      "grad_norm": 0.050981760025024414,
+      "learning_rate": 0.00017024454799658884,
+      "loss": 0.866,
+      "step": 914
+    },
+    {
+      "epoch": 0.51,
+      "grad_norm": 0.045431166887283325,
+      "learning_rate": 0.00017018216774434828,
+      "loss": 0.7997,
+      "step": 915
+    },
+    {
+      "epoch": 0.51,
+      "grad_norm": 0.049921710044145584,
+      "learning_rate": 0.00017011973362719932,
+      "loss": 0.6015,
+      "step": 916
+    },
+    {
+      "epoch": 0.51,
+      "grad_norm": 0.04454491659998894,
+      "learning_rate": 0.00017005724569306026,
+      "loss": 0.6789,
+      "step": 917
+    },
+    {
+      "epoch": 0.51,
+      "grad_norm": 0.04739166796207428,
+      "learning_rate": 0.0001699947039898907,
+      "loss": 0.7727,
+      "step": 918
+    },
+    {
+      "epoch": 0.51,
+      "grad_norm": 0.04357178509235382,
+      "learning_rate": 0.0001699321085656914,
+      "loss": 0.71,
+      "step": 919
+    },
+    {
+      "epoch": 0.51,
+      "grad_norm": 0.04351083189249039,
+      "learning_rate": 0.00016986945946850446,
+      "loss": 0.7437,
+      "step": 920
+    },
+    {
+      "epoch": 0.51,
+      "grad_norm": 0.04492335394024849,
+      "learning_rate": 0.00016980675674641322,
+      "loss": 0.7705,
+      "step": 921
+    },
+    {
+      "epoch": 0.51,
+      "grad_norm": 0.04446355253458023,
+      "learning_rate": 0.000169744000447542,
+      "loss": 0.6758,
+      "step": 922
+    },
+    {
+      "epoch": 0.51,
+      "grad_norm": 0.04474789649248123,
+      "learning_rate": 0.00016968119062005642,
+      "loss": 0.665,
+      "step": 923
+    },
+    {
+      "epoch": 0.51,
+      "grad_norm": 0.03871838003396988,
+      "learning_rate": 0.00016961832731216307,
+      "loss": 0.644,
+      "step": 924
+    },
+    {
+      "epoch": 0.51,
+      "grad_norm": 0.04366254433989525,
+      "learning_rate": 0.00016955541057210965,
+      "loss": 0.8131,
+      "step": 925
+    },
+    {
+      "epoch": 0.52,
+      "grad_norm": 0.04740358516573906,
+      "learning_rate": 0.0001694924404481848,
+      "loss": 0.6964,
+      "step": 926
+    },
+    {
+      "epoch": 0.52,
+      "grad_norm": 0.04698627069592476,
+      "learning_rate": 0.00016942941698871818,
+      "loss": 0.6736,
+      "step": 927
+    },
+    {
+      "epoch": 0.52,
+      "grad_norm": 0.0437115915119648,
+      "learning_rate": 0.00016936634024208047,
+      "loss": 0.7068,
+      "step": 928
+    },
+    {
+      "epoch": 0.52,
+      "grad_norm": 0.04518444836139679,
+      "learning_rate": 0.00016930321025668306,
+      "loss": 0.7357,
+      "step": 929
+    },
+    {
+      "epoch": 0.52,
+      "grad_norm": 0.0465158186852932,
+      "learning_rate": 0.00016924002708097833,
+      "loss": 0.7876,
+      "step": 930
+    },
+    {
+      "epoch": 0.52,
+      "grad_norm": 0.04457319527864456,
+      "learning_rate": 0.00016917679076345943,
+      "loss": 0.7335,
+      "step": 931
+    },
+    {
+      "epoch": 0.52,
+      "grad_norm": 0.05104387551546097,
+      "learning_rate": 0.00016911350135266035,
+      "loss": 0.7055,
+      "step": 932
+    },
+    {
+      "epoch": 0.52,
+      "grad_norm": 0.04391132667660713,
+      "learning_rate": 0.0001690501588971558,
+      "loss": 0.7218,
+      "step": 933
+    },
+    {
+      "epoch": 0.52,
+      "grad_norm": 0.042982205748558044,
+      "learning_rate": 0.00016898676344556118,
+      "loss": 0.7202,
+      "step": 934
+    },
+    {
+      "epoch": 0.52,
+      "grad_norm": 0.042211346328258514,
+      "learning_rate": 0.00016892331504653259,
+      "loss": 0.708,
+      "step": 935
+    },
+    {
+      "epoch": 0.52,
+      "grad_norm": 0.043890226632356644,
+      "learning_rate": 0.00016885981374876677,
+      "loss": 0.7524,
+      "step": 936
+    },
+    {
+      "epoch": 0.52,
+      "grad_norm": 0.044012345373630524,
+      "learning_rate": 0.00016879625960100104,
+      "loss": 0.7103,
+      "step": 937
+    },
+    {
+      "epoch": 0.52,
+      "grad_norm": 0.0452919527888298,
+      "learning_rate": 0.0001687326526520133,
+      "loss": 0.7012,
+      "step": 938
+    },
+    {
+      "epoch": 0.52,
+      "grad_norm": 0.04428261145949364,
+      "learning_rate": 0.00016866899295062197,
+      "loss": 0.6329,
+      "step": 939
+    },
+    {
+      "epoch": 0.52,
+      "grad_norm": 0.043744850903749466,
+      "learning_rate": 0.00016860528054568597,
+      "loss": 0.7682,
+      "step": 940
+    },
+    {
+      "epoch": 0.52,
+      "grad_norm": 0.0442328006029129,
+      "learning_rate": 0.00016854151548610462,
+      "loss": 0.6931,
+      "step": 941
+    },
+    {
+      "epoch": 0.52,
+      "grad_norm": 0.044094718992710114,
+      "learning_rate": 0.00016847769782081772,
+      "loss": 0.7495,
+      "step": 942
+    },
+    {
+      "epoch": 0.52,
+      "grad_norm": 0.041264671832323074,
+      "learning_rate": 0.00016841382759880542,
+      "loss": 0.6035,
+      "step": 943
+    },
+    {
+      "epoch": 0.53,
+      "grad_norm": 0.03848943114280701,
+      "learning_rate": 0.00016834990486908817,
+      "loss": 0.6,
+      "step": 944
+    },
+    {
+      "epoch": 0.53,
+      "grad_norm": 0.0429924838244915,
+      "learning_rate": 0.00016828592968072678,
+      "loss": 0.7149,
+      "step": 945
+    },
+    {
+      "epoch": 0.53,
+      "grad_norm": 0.037183403968811035,
+      "learning_rate": 0.00016822190208282226,
+      "loss": 0.6253,
+      "step": 946
+    },
+    {
+      "epoch": 0.53,
+      "grad_norm": 0.04193463176488876,
+      "learning_rate": 0.00016815782212451592,
+      "loss": 0.7332,
+      "step": 947
+    },
+    {
+      "epoch": 0.53,
+      "grad_norm": 0.04574836045503616,
+      "learning_rate": 0.00016809368985498918,
+      "loss": 0.676,
+      "step": 948
+    },
+    {
+      "epoch": 0.53,
+      "grad_norm": 0.044316504150629044,
+      "learning_rate": 0.0001680295053234637,
+      "loss": 0.6696,
+      "step": 949
+    },
+    {
+      "epoch": 0.53,
+      "grad_norm": 0.04649119824171066,
+      "learning_rate": 0.00016796526857920112,
+      "loss": 0.6853,
+      "step": 950
+    },
+    {
+      "epoch": 0.53,
+      "grad_norm": 0.041028618812561035,
+      "learning_rate": 0.00016790097967150325,
+      "loss": 0.5797,
+      "step": 951
+    },
+    {
+      "epoch": 0.53,
+      "grad_norm": 0.045178357511758804,
+      "learning_rate": 0.00016783663864971193,
+      "loss": 0.733,
+      "step": 952
+    },
+    {
+      "epoch": 0.53,
+      "grad_norm": 0.044117413461208344,
+      "learning_rate": 0.00016777224556320896,
+      "loss": 0.6571,
+      "step": 953
+    },
+    {
+      "epoch": 0.53,
+      "grad_norm": 0.04647046700119972,
+      "learning_rate": 0.00016770780046141616,
+      "loss": 0.7674,
+      "step": 954
+    },
+    {
+      "epoch": 0.53,
+      "grad_norm": 0.04585673660039902,
+      "learning_rate": 0.0001676433033937952,
+      "loss": 0.7153,
+      "step": 955
+    },
+    {
+      "epoch": 0.53,
+      "grad_norm": 0.04684153199195862,
+      "learning_rate": 0.00016757875440984768,
+      "loss": 0.712,
+      "step": 956
+    },
+    {
+      "epoch": 0.53,
+      "grad_norm": 0.05014558881521225,
+      "learning_rate": 0.000167514153559115,
+      "loss": 0.7872,
+      "step": 957
+    },
+    {
+      "epoch": 0.53,
+      "grad_norm": 0.04870294779539108,
+      "learning_rate": 0.00016744950089117846,
+      "loss": 0.7779,
+      "step": 958
+    },
+    {
+      "epoch": 0.53,
+      "grad_norm": 0.047236453741788864,
+      "learning_rate": 0.00016738479645565902,
+      "loss": 0.6254,
+      "step": 959
+    },
+    {
+      "epoch": 0.53,
+      "grad_norm": 0.04619355872273445,
+      "learning_rate": 0.00016732004030221745,
+      "loss": 0.7061,
+      "step": 960
+    },
+    {
+      "epoch": 0.53,
+      "grad_norm": 0.04691663756966591,
+      "learning_rate": 0.00016725523248055415,
+      "loss": 0.7212,
+      "step": 961
+    },
+    {
+      "epoch": 0.54,
+      "grad_norm": 0.046586498618125916,
+      "learning_rate": 0.00016719037304040922,
+      "loss": 0.7438,
+      "step": 962
+    },
+    {
+      "epoch": 0.54,
+      "grad_norm": 0.047182347625494,
+      "learning_rate": 0.00016712546203156243,
+      "loss": 0.7781,
+      "step": 963
+    },
+    {
+      "epoch": 0.54,
+      "grad_norm": 0.04682426154613495,
+      "learning_rate": 0.000167060499503833,
+      "loss": 0.7651,
+      "step": 964
+    },
+    {
+      "epoch": 0.54,
+      "grad_norm": 0.04150580242276192,
+      "learning_rate": 0.00016699548550707974,
+      "loss": 0.6261,
+      "step": 965
+    },
+    {
+      "epoch": 0.54,
+      "grad_norm": 0.047207970172166824,
+      "learning_rate": 0.00016693042009120106,
+      "loss": 0.7482,
+      "step": 966
+    },
+    {
+      "epoch": 0.54,
+      "grad_norm": 0.03972497582435608,
+      "learning_rate": 0.0001668653033061347,
+      "loss": 0.6568,
+      "step": 967
+    },
+    {
+      "epoch": 0.54,
+      "grad_norm": 0.05301308631896973,
+      "learning_rate": 0.00016680013520185786,
+      "loss": 0.8632,
+      "step": 968
+    },
+    {
+      "epoch": 0.54,
+      "grad_norm": 0.044552650302648544,
+      "learning_rate": 0.0001667349158283872,
+      "loss": 0.7905,
+      "step": 969
+    },
+    {
+      "epoch": 0.54,
+      "grad_norm": 0.0502367801964283,
+      "learning_rate": 0.0001666696452357787,
+      "loss": 0.694,
+      "step": 970
+    },
+    {
+      "epoch": 0.54,
+      "grad_norm": 0.047164347022771835,
+      "learning_rate": 0.00016660432347412752,
+      "loss": 0.6664,
+      "step": 971
+    },
+    {
+      "epoch": 0.54,
+      "grad_norm": 0.04584898054599762,
+      "learning_rate": 0.00016653895059356828,
+      "loss": 0.6614,
+      "step": 972
+    },
+    {
+      "epoch": 0.54,
+      "grad_norm": 0.04450858384370804,
+      "learning_rate": 0.00016647352664427473,
+      "loss": 0.7293,
+      "step": 973
+    },
+    {
+      "epoch": 0.54,
+      "grad_norm": 0.045772675424814224,
+      "learning_rate": 0.00016640805167645985,
+      "loss": 0.6825,
+      "step": 974
+    },
+    {
+      "epoch": 0.54,
+      "grad_norm": 0.04510961472988129,
+      "learning_rate": 0.00016634252574037575,
+      "loss": 0.628,
+      "step": 975
+    },
+    {
+      "epoch": 0.54,
+      "grad_norm": 0.0424661748111248,
+      "learning_rate": 0.00016627694888631377,
+      "loss": 0.7242,
+      "step": 976
+    },
+    {
+      "epoch": 0.54,
+      "grad_norm": 0.04516896605491638,
+      "learning_rate": 0.00016621132116460404,
+      "loss": 0.728,
+      "step": 977
+    },
+    {
+      "epoch": 0.54,
+      "grad_norm": 0.04178481549024582,
+      "learning_rate": 0.00016614564262561608,
+      "loss": 0.6749,
+      "step": 978
+    },
+    {
+      "epoch": 0.54,
+      "grad_norm": 0.04447667673230171,
+      "learning_rate": 0.00016607991331975822,
+      "loss": 0.7377,
+      "step": 979
+    },
+    {
+      "epoch": 0.55,
+      "grad_norm": 0.04504280164837837,
+      "learning_rate": 0.0001660141332974778,
+      "loss": 0.6817,
+      "step": 980
+    },
+    {
+      "epoch": 0.55,
+      "grad_norm": 0.0446137972176075,
+      "learning_rate": 0.00016594830260926102,
+      "loss": 0.7593,
+      "step": 981
+    },
+    {
+      "epoch": 0.55,
+      "grad_norm": 0.04577165096998215,
+      "learning_rate": 0.0001658824213056331,
+      "loss": 0.6594,
+      "step": 982
+    },
+    {
+      "epoch": 0.55,
+      "grad_norm": 0.04543627053499222,
+      "learning_rate": 0.00016581648943715796,
+      "loss": 0.7227,
+      "step": 983
+    },
+    {
+      "epoch": 0.55,
+      "grad_norm": 0.051089901477098465,
+      "learning_rate": 0.0001657505070544384,
+      "loss": 0.6931,
+      "step": 984
+    },
+    {
+      "epoch": 0.55,
+      "grad_norm": 0.04433685168623924,
+      "learning_rate": 0.00016568447420811612,
+      "loss": 0.7223,
+      "step": 985
+    },
+    {
+      "epoch": 0.55,
+      "grad_norm": 0.04630895331501961,
+      "learning_rate": 0.00016561839094887123,
+      "loss": 0.8788,
+      "step": 986
+    },
+    {
+      "epoch": 0.55,
+      "grad_norm": 0.04735618084669113,
+      "learning_rate": 0.00016555225732742281,
+      "loss": 0.6527,
+      "step": 987
+    },
+    {
+      "epoch": 0.55,
+      "grad_norm": 0.04732111841440201,
+      "learning_rate": 0.00016548607339452853,
+      "loss": 0.7427,
+      "step": 988
+    },
+    {
+      "epoch": 0.55,
+      "grad_norm": 0.04409315809607506,
+      "learning_rate": 0.0001654198392009846,
+      "loss": 0.7473,
+      "step": 989
+    },
+    {
+      "epoch": 0.55,
+      "grad_norm": 0.044037654995918274,
+      "learning_rate": 0.00016535355479762586,
+      "loss": 0.735,
+      "step": 990
+    },
+    {
+      "epoch": 0.55,
+      "grad_norm": 0.043528955429792404,
+      "learning_rate": 0.00016528722023532574,
+      "loss": 0.6415,
+      "step": 991
+    },
+    {
+      "epoch": 0.55,
+      "grad_norm": 0.04991381987929344,
+      "learning_rate": 0.00016522083556499596,
+      "loss": 0.7989,
+      "step": 992
+    },
+    {
+      "epoch": 0.55,
+      "grad_norm": 0.044087547808885574,
+      "learning_rate": 0.000165154400837587,
+      "loss": 0.7178,
+      "step": 993
+    },
+    {
+      "epoch": 0.55,
+      "grad_norm": 0.04327988997101784,
+      "learning_rate": 0.00016508791610408754,
+      "loss": 0.6599,
+      "step": 994
+    },
+    {
+      "epoch": 0.55,
+      "grad_norm": 0.04584032669663429,
+      "learning_rate": 0.0001650213814155247,
+      "loss": 0.6482,
+      "step": 995
+    },
+    {
+      "epoch": 0.55,
+      "grad_norm": 0.05077075958251953,
+      "learning_rate": 0.00016495479682296395,
+      "loss": 0.727,
+      "step": 996
+    },
+    {
+      "epoch": 0.55,
+      "grad_norm": 0.04770886152982712,
+      "learning_rate": 0.00016488816237750906,
+      "loss": 0.6737,
+      "step": 997
+    },
+    {
+      "epoch": 0.56,
+      "grad_norm": 0.04274902865290642,
+      "learning_rate": 0.00016482147813030203,
+      "loss": 0.6277,
+      "step": 998
+    },
+    {
+      "epoch": 0.56,
+      "grad_norm": 0.048387110233306885,
+      "learning_rate": 0.00016475474413252315,
+      "loss": 0.7836,
+      "step": 999
+    },
+    {
+      "epoch": 0.56,
+      "grad_norm": 0.04784883186221123,
+      "learning_rate": 0.0001646879604353908,
+      "loss": 0.8291,
+      "step": 1000
+    },
+    {
+      "epoch": 0.56,
+      "grad_norm": 0.047025181353092194,
+      "learning_rate": 0.00016462112709016163,
+      "loss": 0.7344,
+      "step": 1001
+    },
+    {
+      "epoch": 0.56,
+      "grad_norm": 0.042724139988422394,
+      "learning_rate": 0.00016455424414813026,
+      "loss": 0.607,
+      "step": 1002
+    },
+    {
+      "epoch": 0.56,
+      "grad_norm": 0.04707411676645279,
+      "learning_rate": 0.00016448731166062946,
+      "loss": 0.7879,
+      "step": 1003
+    },
+    {
+      "epoch": 0.56,
+      "grad_norm": 0.048031605780124664,
+      "learning_rate": 0.00016442032967903,
+      "loss": 0.7665,
+      "step": 1004
+    },
+    {
+      "epoch": 0.56,
+      "grad_norm": 0.0522460862994194,
+      "learning_rate": 0.00016435329825474065,
+      "loss": 0.8697,
+      "step": 1005
+    },
+    {
+      "epoch": 0.56,
+      "grad_norm": 0.043403007090091705,
+      "learning_rate": 0.00016428621743920814,
+      "loss": 0.6733,
+      "step": 1006
+    },
+    {
+      "epoch": 0.56,
+      "grad_norm": 0.04464132711291313,
+      "learning_rate": 0.00016421908728391703,
+      "loss": 0.6495,
+      "step": 1007
+    },
+    {
+      "epoch": 0.56,
+      "grad_norm": 0.04746426269412041,
+      "learning_rate": 0.00016415190784038983,
+      "loss": 0.6364,
+      "step": 1008
+    },
+    {
+      "epoch": 0.56,
+      "grad_norm": 0.04561330005526543,
+      "learning_rate": 0.00016408467916018688,
+      "loss": 0.7492,
+      "step": 1009
+    },
+    {
+      "epoch": 0.56,
+      "grad_norm": 0.046681590378284454,
+      "learning_rate": 0.00016401740129490624,
+      "loss": 0.6807,
+      "step": 1010
+    },
+    {
+      "epoch": 0.56,
+      "grad_norm": 0.04807223007082939,
+      "learning_rate": 0.00016395007429618382,
+      "loss": 0.7732,
+      "step": 1011
+    },
+    {
+      "epoch": 0.56,
+      "grad_norm": 0.04678153246641159,
+      "learning_rate": 0.00016388269821569312,
+      "loss": 0.753,
+      "step": 1012
+    },
+    {
+      "epoch": 0.56,
+      "grad_norm": 0.0424916073679924,
+      "learning_rate": 0.0001638152731051454,
+      "loss": 0.6378,
+      "step": 1013
+    },
+    {
+      "epoch": 0.56,
+      "grad_norm": 0.04331868886947632,
+      "learning_rate": 0.00016374779901628952,
+      "loss": 0.766,
+      "step": 1014
+    },
+    {
+      "epoch": 0.56,
+      "grad_norm": 0.043369825929403305,
+      "learning_rate": 0.00016368027600091194,
+      "loss": 0.6492,
+      "step": 1015
+    },
+    {
+      "epoch": 0.57,
+      "grad_norm": 0.04750450327992439,
+      "learning_rate": 0.00016361270411083666,
+      "loss": 0.7226,
+      "step": 1016
+    },
+    {
+      "epoch": 0.57,
+      "grad_norm": 0.04926105588674545,
+      "learning_rate": 0.0001635450833979252,
+      "loss": 0.7402,
+      "step": 1017
+    },
+    {
+      "epoch": 0.57,
+      "grad_norm": 0.04888901859521866,
+      "learning_rate": 0.00016347741391407655,
+      "loss": 0.7061,
+      "step": 1018
+    },
+    {
+      "epoch": 0.57,
+      "grad_norm": 0.046233415603637695,
+      "learning_rate": 0.00016340969571122712,
+      "loss": 0.7367,
+      "step": 1019
+    },
+    {
+      "epoch": 0.57,
+      "grad_norm": 0.0504348948597908,
+      "learning_rate": 0.00016334192884135074,
+      "loss": 0.7785,
+      "step": 1020
+    },
+    {
+      "epoch": 0.57,
+      "grad_norm": 0.04318249598145485,
+      "learning_rate": 0.00016327411335645853,
+      "loss": 0.6495,
+      "step": 1021
+    },
+    {
+      "epoch": 0.57,
+      "grad_norm": 0.04340077564120293,
+      "learning_rate": 0.00016320624930859904,
+      "loss": 0.6685,
+      "step": 1022
+    },
+    {
+      "epoch": 0.57,
+      "grad_norm": 0.04627622663974762,
+      "learning_rate": 0.00016313833674985796,
+      "loss": 0.6857,
+      "step": 1023
+    },
+    {
+      "epoch": 0.57,
+      "grad_norm": 0.04838955029845238,
+      "learning_rate": 0.0001630703757323583,
+      "loss": 0.7336,
+      "step": 1024
+    },
+    {
+      "epoch": 0.57,
+      "grad_norm": 0.04140018671751022,
+      "learning_rate": 0.00016300236630826024,
+      "loss": 0.6872,
+      "step": 1025
+    },
+    {
+      "epoch": 0.57,
+      "grad_norm": 0.04307800903916359,
+      "learning_rate": 0.000162934308529761,
+      "loss": 0.689,
+      "step": 1026
+    },
+    {
+      "epoch": 0.57,
+      "grad_norm": 0.046013642102479935,
+      "learning_rate": 0.00016286620244909518,
+      "loss": 0.676,
+      "step": 1027
+    },
+    {
+      "epoch": 0.57,
+      "grad_norm": 0.045836132019758224,
+      "learning_rate": 0.0001627980481185342,
+      "loss": 0.7321,
+      "step": 1028
+    },
+    {
+      "epoch": 0.57,
+      "grad_norm": 0.046595752239227295,
+      "learning_rate": 0.00016272984559038655,
+      "loss": 0.7788,
+      "step": 1029
+    },
+    {
+      "epoch": 0.57,
+      "grad_norm": 0.041627172380685806,
+      "learning_rate": 0.0001626615949169979,
+      "loss": 0.5976,
+      "step": 1030
+    },
+    {
+      "epoch": 0.57,
+      "grad_norm": 0.056944046169519424,
+      "learning_rate": 0.00016259329615075062,
+      "loss": 0.937,
+      "step": 1031
+    },
+    {
+      "epoch": 0.57,
+      "grad_norm": 0.0483165942132473,
+      "learning_rate": 0.00016252494934406415,
+      "loss": 0.6425,
+      "step": 1032
+    },
+    {
+      "epoch": 0.57,
+      "grad_norm": 0.05059583857655525,
+      "learning_rate": 0.00016245655454939474,
+      "loss": 0.7802,
+      "step": 1033
+    },
+    {
+      "epoch": 0.58,
+      "grad_norm": 0.04494575038552284,
+      "learning_rate": 0.0001623881118192355,
+      "loss": 0.709,
+      "step": 1034
+    },
+    {
+      "epoch": 0.58,
+      "grad_norm": 0.04515180364251137,
+      "learning_rate": 0.00016231962120611635,
+      "loss": 0.7219,
+      "step": 1035
+    },
+    {
+      "epoch": 0.58,
+      "grad_norm": 0.051081229001283646,
+      "learning_rate": 0.00016225108276260385,
+      "loss": 0.7955,
+      "step": 1036
+    },
+    {
+      "epoch": 0.58,
+      "grad_norm": 0.04739641770720482,
+      "learning_rate": 0.00016218249654130138,
+      "loss": 0.7017,
+      "step": 1037
+    },
+    {
+      "epoch": 0.58,
+      "grad_norm": 0.048389632254838943,
+      "learning_rate": 0.000162113862594849,
+      "loss": 0.7864,
+      "step": 1038
+    },
+    {
+      "epoch": 0.58,
+      "grad_norm": 0.04880169779062271,
+      "learning_rate": 0.0001620451809759233,
+      "loss": 0.7688,
+      "step": 1039
+    },
+    {
+      "epoch": 0.58,
+      "grad_norm": 0.044919900596141815,
+      "learning_rate": 0.00016197645173723758,
+      "loss": 0.754,
+      "step": 1040
+    },
+    {
+      "epoch": 0.58,
+      "grad_norm": 0.04826303571462631,
+      "learning_rate": 0.00016190767493154154,
+      "loss": 0.7931,
+      "step": 1041
+    },
+    {
+      "epoch": 0.58,
+      "grad_norm": 0.042201928794384,
+      "learning_rate": 0.0001618388506116215,
+      "loss": 0.603,
+      "step": 1042
+    },
+    {
+      "epoch": 0.58,
+      "grad_norm": 0.051390066742897034,
+      "learning_rate": 0.00016176997883030026,
+      "loss": 0.7412,
+      "step": 1043
+    },
+    {
+      "epoch": 0.58,
+      "grad_norm": 0.045629408210515976,
+      "learning_rate": 0.00016170105964043695,
+      "loss": 0.7082,
+      "step": 1044
+    },
+    {
+      "epoch": 0.58,
+      "grad_norm": 0.04842764884233475,
+      "learning_rate": 0.0001616320930949272,
+      "loss": 0.7796,
+      "step": 1045
+    },
+    {
+      "epoch": 0.58,
+      "grad_norm": 0.04266642406582832,
+      "learning_rate": 0.00016156307924670288,
+      "loss": 0.6521,
+      "step": 1046
+    },
+    {
+      "epoch": 0.58,
+      "grad_norm": 0.04461183398962021,
+      "learning_rate": 0.00016149401814873223,
+      "loss": 0.7277,
+      "step": 1047
+    },
+    {
+      "epoch": 0.58,
+      "grad_norm": 0.044011376798152924,
+      "learning_rate": 0.0001614249098540197,
+      "loss": 0.7393,
+      "step": 1048
+    },
+    {
+      "epoch": 0.58,
+      "grad_norm": 0.04548850655555725,
+      "learning_rate": 0.0001613557544156061,
+      "loss": 0.6724,
+      "step": 1049
+    },
+    {
+      "epoch": 0.58,
+      "grad_norm": 0.04711510241031647,
+      "learning_rate": 0.00016128655188656817,
+      "loss": 0.75,
+      "step": 1050
+    },
+    {
+      "epoch": 0.58,
+      "grad_norm": 0.04443375766277313,
+      "learning_rate": 0.00016121730232001905,
+      "loss": 0.692,
+      "step": 1051
+    },
+    {
+      "epoch": 0.59,
+      "grad_norm": 0.047497253865003586,
+      "learning_rate": 0.00016114800576910788,
+      "loss": 0.6364,
+      "step": 1052
+    },
+    {
+      "epoch": 0.59,
+      "grad_norm": 0.04716882482171059,
+      "learning_rate": 0.00016107866228701981,
+      "loss": 0.8157,
+      "step": 1053
+    },
+    {
+      "epoch": 0.59,
+      "grad_norm": 0.049585748463869095,
+      "learning_rate": 0.0001610092719269761,
+      "loss": 0.7222,
+      "step": 1054
+    },
+    {
+      "epoch": 0.59,
+      "grad_norm": 0.04663019999861717,
+      "learning_rate": 0.0001609398347422339,
+      "loss": 0.674,
+      "step": 1055
+    },
+    {
+      "epoch": 0.59,
+      "grad_norm": 0.04515184462070465,
+      "learning_rate": 0.00016087035078608637,
+      "loss": 0.641,
+      "step": 1056
+    },
+    {
+      "epoch": 0.59,
+      "grad_norm": 0.04762982949614525,
+      "learning_rate": 0.0001608008201118625,
+      "loss": 0.6312,
+      "step": 1057
+    },
+    {
+      "epoch": 0.59,
+      "grad_norm": 0.04451216757297516,
+      "learning_rate": 0.00016073124277292728,
+      "loss": 0.7158,
+      "step": 1058
+    },
+    {
+      "epoch": 0.59,
+      "grad_norm": 0.0480036623775959,
+      "learning_rate": 0.0001606616188226813,
+      "loss": 0.7029,
+      "step": 1059
+    },
+    {
+      "epoch": 0.59,
+      "grad_norm": 0.04552299529314041,
+      "learning_rate": 0.00016059194831456107,
+      "loss": 0.7185,
+      "step": 1060
+    },
+    {
+      "epoch": 0.59,
+      "grad_norm": 0.04811175912618637,
+      "learning_rate": 0.0001605222313020388,
+      "loss": 0.5746,
+      "step": 1061
+    },
+    {
+      "epoch": 0.59,
+      "grad_norm": 0.045087020844221115,
+      "learning_rate": 0.0001604524678386224,
+      "loss": 0.7196,
+      "step": 1062
+    },
+    {
+      "epoch": 0.59,
+      "grad_norm": 0.05363542214035988,
+      "learning_rate": 0.00016038265797785542,
+      "loss": 0.7144,
+      "step": 1063
+    },
+    {
+      "epoch": 0.59,
+      "grad_norm": 0.05288343131542206,
+      "learning_rate": 0.00016031280177331706,
+      "loss": 0.7173,
+      "step": 1064
+    },
+    {
+      "epoch": 0.59,
+      "grad_norm": 0.046650927513837814,
+      "learning_rate": 0.00016024289927862195,
+      "loss": 0.789,
+      "step": 1065
+    },
+    {
+      "epoch": 0.59,
+      "grad_norm": 0.048903193324804306,
+      "learning_rate": 0.00016017295054742046,
+      "loss": 0.7106,
+      "step": 1066
+    },
+    {
+      "epoch": 0.59,
+      "grad_norm": 0.04834052920341492,
+      "learning_rate": 0.00016010295563339825,
+      "loss": 0.6996,
+      "step": 1067
+    },
+    {
+      "epoch": 0.59,
+      "grad_norm": 0.05408628657460213,
+      "learning_rate": 0.00016003291459027653,
+      "loss": 0.8056,
+      "step": 1068
+    },
+    {
+      "epoch": 0.59,
+      "grad_norm": 0.04541575536131859,
+      "learning_rate": 0.00015996282747181197,
+      "loss": 0.6443,
+      "step": 1069
+    },
+    {
+      "epoch": 0.6,
+      "grad_norm": 0.04552549123764038,
+      "learning_rate": 0.00015989269433179645,
+      "loss": 0.69,
+      "step": 1070
+    },
+    {
+      "epoch": 0.6,
+      "grad_norm": 0.0493139810860157,
+      "learning_rate": 0.00015982251522405727,
+      "loss": 0.7327,
+      "step": 1071
+    },
+    {
+      "epoch": 0.6,
+      "grad_norm": 0.049078866839408875,
+      "learning_rate": 0.000159752290202457,
+      "loss": 0.8271,
+      "step": 1072
+    },
+    {
+      "epoch": 0.6,
+      "grad_norm": 0.0488913394510746,
+      "learning_rate": 0.0001596820193208934,
+      "loss": 0.7298,
+      "step": 1073
+    },
+    {
+      "epoch": 0.6,
+      "grad_norm": 0.04880421981215477,
+      "learning_rate": 0.00015961170263329948,
+      "loss": 0.6863,
+      "step": 1074
+    },
+    {
+      "epoch": 0.6,
+      "grad_norm": 0.04644745960831642,
+      "learning_rate": 0.00015954134019364346,
+      "loss": 0.6754,
+      "step": 1075
+    },
+    {
+      "epoch": 0.6,
+      "grad_norm": 0.05141911655664444,
+      "learning_rate": 0.00015947093205592855,
+      "loss": 0.7986,
+      "step": 1076
+    },
+    {
+      "epoch": 0.6,
+      "grad_norm": 0.052950434386730194,
+      "learning_rate": 0.00015940047827419303,
+      "loss": 0.8126,
+      "step": 1077
+    },
+    {
+      "epoch": 0.6,
+      "grad_norm": 0.04494583606719971,
+      "learning_rate": 0.0001593299789025104,
+      "loss": 0.724,
+      "step": 1078
+    },
+    {
+      "epoch": 0.6,
+      "grad_norm": 0.051292311400175095,
+      "learning_rate": 0.00015925943399498898,
+      "loss": 0.7201,
+      "step": 1079
+    },
+    {
+      "epoch": 0.6,
+      "grad_norm": 0.049635812640190125,
+      "learning_rate": 0.000159188843605772,
+      "loss": 0.6293,
+      "step": 1080
+    },
+    {
+      "epoch": 0.6,
+      "grad_norm": 0.08753520250320435,
+      "learning_rate": 0.00015911820778903777,
+      "loss": 0.7986,
+      "step": 1081
+    },
+    {
+      "epoch": 0.6,
+      "grad_norm": 0.05015747994184494,
+      "learning_rate": 0.00015904752659899935,
+      "loss": 0.7774,
+      "step": 1082
+    },
+    {
+      "epoch": 0.6,
+      "grad_norm": 0.05062435567378998,
+      "learning_rate": 0.00015897680008990467,
+      "loss": 0.7423,
+      "step": 1083
+    },
+    {
+      "epoch": 0.6,
+      "grad_norm": 0.04625916853547096,
+      "learning_rate": 0.00015890602831603634,
+      "loss": 0.7179,
+      "step": 1084
+    },
+    {
+      "epoch": 0.6,
+      "grad_norm": 0.06423801183700562,
+      "learning_rate": 0.00015883521133171186,
+      "loss": 0.6282,
+      "step": 1085
+    },
+    {
+      "epoch": 0.6,
+      "grad_norm": 0.048834312707185745,
+      "learning_rate": 0.00015876434919128335,
+      "loss": 0.7035,
+      "step": 1086
+    },
+    {
+      "epoch": 0.6,
+      "grad_norm": 0.049968525767326355,
+      "learning_rate": 0.0001586934419491376,
+      "loss": 0.777,
+      "step": 1087
+    },
+    {
+      "epoch": 0.61,
+      "grad_norm": 0.04475943744182587,
+      "learning_rate": 0.00015862248965969604,
+      "loss": 0.6788,
+      "step": 1088
+    },
+    {
+      "epoch": 0.61,
+      "grad_norm": 0.046110913157463074,
+      "learning_rate": 0.0001585514923774146,
+      "loss": 0.7237,
+      "step": 1089
+    },
+    {
+      "epoch": 0.61,
+      "grad_norm": 0.047238875180482864,
+      "learning_rate": 0.0001584804501567838,
+      "loss": 0.7412,
+      "step": 1090
+    },
+    {
+      "epoch": 0.61,
+      "grad_norm": 0.04489819332957268,
+      "learning_rate": 0.00015840936305232869,
+      "loss": 0.6327,
+      "step": 1091
+    },
+    {
+      "epoch": 0.61,
+      "grad_norm": 0.04712117463350296,
+      "learning_rate": 0.00015833823111860863,
+      "loss": 0.7175,
+      "step": 1092
+    },
+    {
+      "epoch": 0.61,
+      "grad_norm": 0.05276874080300331,
+      "learning_rate": 0.0001582670544102175,
+      "loss": 0.626,
+      "step": 1093
+    },
+    {
+      "epoch": 0.61,
+      "grad_norm": 0.047401316463947296,
+      "learning_rate": 0.0001581958329817836,
+      "loss": 0.7426,
+      "step": 1094
+    },
+    {
+      "epoch": 0.61,
+      "grad_norm": 0.049675095826387405,
+      "learning_rate": 0.0001581245668879694,
+      "loss": 0.7332,
+      "step": 1095
+    },
+    {
+      "epoch": 0.61,
+      "grad_norm": 0.048146385699510574,
+      "learning_rate": 0.00015805325618347172,
+      "loss": 0.7209,
+      "step": 1096
+    },
+    {
+      "epoch": 0.61,
+      "grad_norm": 0.04951384291052818,
+      "learning_rate": 0.00015798190092302164,
+      "loss": 0.6746,
+      "step": 1097
+    },
+    {
+      "epoch": 0.61,
+      "grad_norm": 0.0463559590280056,
+      "learning_rate": 0.0001579105011613844,
+      "loss": 0.6808,
+      "step": 1098
+    },
+    {
+      "epoch": 0.61,
+      "grad_norm": 0.04673503711819649,
+      "learning_rate": 0.00015783905695335946,
+      "loss": 0.6946,
+      "step": 1099
+    },
+    {
+      "epoch": 0.61,
+      "grad_norm": 0.04709373414516449,
+      "learning_rate": 0.0001577675683537803,
+      "loss": 0.691,
+      "step": 1100
+    },
+    {
+      "epoch": 0.61,
+      "grad_norm": 0.05507972836494446,
+      "learning_rate": 0.00015769603541751455,
+      "loss": 0.7342,
+      "step": 1101
+    },
+    {
+      "epoch": 0.61,
+      "grad_norm": 0.043717097491025925,
+      "learning_rate": 0.00015762445819946384,
+      "loss": 0.6183,
+      "step": 1102
+    },
+    {
+      "epoch": 0.61,
+      "grad_norm": 0.04856884479522705,
+      "learning_rate": 0.0001575528367545637,
+      "loss": 0.7489,
+      "step": 1103
+    },
+    {
+      "epoch": 0.61,
+      "grad_norm": 0.04887085407972336,
+      "learning_rate": 0.0001574811711377838,
+      "loss": 0.7202,
+      "step": 1104
+    },
+    {
+      "epoch": 0.61,
+      "grad_norm": 0.05098757520318031,
+      "learning_rate": 0.00015740946140412753,
+      "loss": 0.7171,
+      "step": 1105
+    },
+    {
+      "epoch": 0.62,
+      "grad_norm": 0.05076475441455841,
+      "learning_rate": 0.0001573377076086322,
+      "loss": 0.6581,
+      "step": 1106
+    },
+    {
+      "epoch": 0.62,
+      "grad_norm": 0.05398886650800705,
+      "learning_rate": 0.00015726590980636896,
+      "loss": 0.8532,
+      "step": 1107
+    },
+    {
+      "epoch": 0.62,
+      "grad_norm": 0.05237511545419693,
+      "learning_rate": 0.00015719406805244276,
+      "loss": 0.6418,
+      "step": 1108
+    },
+    {
+      "epoch": 0.62,
+      "grad_norm": 0.05127400532364845,
+      "learning_rate": 0.00015712218240199218,
+      "loss": 0.75,
+      "step": 1109
+    },
+    {
+      "epoch": 0.62,
+      "grad_norm": 0.051004305481910706,
+      "learning_rate": 0.0001570502529101896,
+      "loss": 0.6545,
+      "step": 1110
+    },
+    {
+      "epoch": 0.62,
+      "grad_norm": 0.05179214105010033,
+      "learning_rate": 0.000156978279632241,
+      "loss": 0.6854,
+      "step": 1111
+    },
+    {
+      "epoch": 0.62,
+      "grad_norm": 0.04017964377999306,
+      "learning_rate": 0.0001569062626233859,
+      "loss": 0.5733,
+      "step": 1112
+    },
+    {
+      "epoch": 0.62,
+      "grad_norm": 0.05255338177084923,
+      "learning_rate": 0.00015683420193889753,
+      "loss": 0.9239,
+      "step": 1113
+    },
+    {
+      "epoch": 0.62,
+      "grad_norm": 0.04868584871292114,
+      "learning_rate": 0.00015676209763408253,
+      "loss": 0.7038,
+      "step": 1114
+    },
+    {
+      "epoch": 0.62,
+      "grad_norm": 0.04898803308606148,
+      "learning_rate": 0.0001566899497642811,
+      "loss": 0.6677,
+      "step": 1115
+    },
+    {
+      "epoch": 0.62,
+      "grad_norm": 0.044663190841674805,
+      "learning_rate": 0.00015661775838486674,
+      "loss": 0.741,
+      "step": 1116
+    },
+    {
+      "epoch": 0.62,
+      "grad_norm": 0.048298414796590805,
+      "learning_rate": 0.00015654552355124648,
+      "loss": 0.7292,
+      "step": 1117
+    },
+    {
+      "epoch": 0.62,
+      "grad_norm": 0.04956801235675812,
+      "learning_rate": 0.00015647324531886065,
+      "loss": 0.7475,
+      "step": 1118
+    },
+    {
+      "epoch": 0.62,
+      "grad_norm": 0.051773227751255035,
+      "learning_rate": 0.0001564009237431829,
+      "loss": 0.751,
+      "step": 1119
+    },
+    {
+      "epoch": 0.62,
+      "grad_norm": 0.04530036076903343,
+      "learning_rate": 0.00015632855887972008,
+      "loss": 0.7123,
+      "step": 1120
+    },
+    {
+      "epoch": 0.62,
+      "grad_norm": 0.05154275894165039,
+      "learning_rate": 0.00015625615078401244,
+      "loss": 0.8531,
+      "step": 1121
+    },
+    {
+      "epoch": 0.62,
+      "grad_norm": 0.04365009814500809,
+      "learning_rate": 0.00015618369951163317,
+      "loss": 0.6508,
+      "step": 1122
+    },
+    {
+      "epoch": 0.62,
+      "grad_norm": 0.04828101769089699,
+      "learning_rate": 0.00015611120511818877,
+      "loss": 0.6855,
+      "step": 1123
+    },
+    {
+      "epoch": 0.63,
+      "grad_norm": 0.04952272027730942,
+      "learning_rate": 0.00015603866765931875,
+      "loss": 0.6806,
+      "step": 1124
+    },
+    {
+      "epoch": 0.63,
+      "grad_norm": 0.0446164608001709,
+      "learning_rate": 0.00015596608719069578,
+      "loss": 0.6364,
+      "step": 1125
+    },
+    {
+      "epoch": 0.63,
+      "grad_norm": 0.05198860913515091,
+      "learning_rate": 0.00015589346376802544,
+      "loss": 0.6415,
+      "step": 1126
+    },
+    {
+      "epoch": 0.63,
+      "grad_norm": 0.04873516410589218,
+      "learning_rate": 0.00015582079744704626,
+      "loss": 0.6668,
+      "step": 1127
+    },
+    {
+      "epoch": 0.63,
+      "grad_norm": 0.04862043261528015,
+      "learning_rate": 0.00015574808828352978,
+      "loss": 0.6277,
+      "step": 1128
+    },
+    {
+      "epoch": 0.63,
+      "grad_norm": 0.050693485885858536,
+      "learning_rate": 0.00015567533633328035,
+      "loss": 0.6917,
+      "step": 1129
+    },
+    {
+      "epoch": 0.63,
+      "grad_norm": 0.048907842487096786,
+      "learning_rate": 0.00015560254165213522,
+      "loss": 0.7091,
+      "step": 1130
+    },
+    {
+      "epoch": 0.63,
+      "grad_norm": 0.04532686248421669,
+      "learning_rate": 0.00015552970429596438,
+      "loss": 0.622,
+      "step": 1131
+    },
+    {
+      "epoch": 0.63,
+      "grad_norm": 0.043171364814043045,
+      "learning_rate": 0.00015545682432067067,
+      "loss": 0.6719,
+      "step": 1132
+    },
+    {
+      "epoch": 0.63,
+      "grad_norm": 0.05676614120602608,
+      "learning_rate": 0.00015538390178218946,
+      "loss": 0.8366,
+      "step": 1133
+    },
+    {
+      "epoch": 0.63,
+      "grad_norm": 0.045390404760837555,
+      "learning_rate": 0.00015531093673648897,
+      "loss": 0.6162,
+      "step": 1134
+    },
+    {
+      "epoch": 0.63,
+      "grad_norm": 0.05167704075574875,
+      "learning_rate": 0.00015523792923957,
+      "loss": 0.7946,
+      "step": 1135
+    },
+    {
+      "epoch": 0.63,
+      "grad_norm": 0.04950125515460968,
+      "learning_rate": 0.00015516487934746578,
+      "loss": 0.6541,
+      "step": 1136
+    },
+    {
+      "epoch": 0.63,
+      "grad_norm": 0.0524483397603035,
+      "learning_rate": 0.00015509178711624233,
+      "loss": 0.7793,
+      "step": 1137
+    },
+    {
+      "epoch": 0.63,
+      "grad_norm": 0.046446293592453,
+      "learning_rate": 0.00015501865260199795,
+      "loss": 0.6489,
+      "step": 1138
+    },
+    {
+      "epoch": 0.63,
+      "grad_norm": 0.0487704761326313,
+      "learning_rate": 0.00015494547586086352,
+      "loss": 0.6154,
+      "step": 1139
+    },
+    {
+      "epoch": 0.63,
+      "grad_norm": 0.04823431745171547,
+      "learning_rate": 0.00015487225694900224,
+      "loss": 0.6899,
+      "step": 1140
+    },
+    {
+      "epoch": 0.63,
+      "grad_norm": 0.04303416609764099,
+      "learning_rate": 0.00015479899592260983,
+      "loss": 0.5816,
+      "step": 1141
+    },
+    {
+      "epoch": 0.64,
+      "grad_norm": 0.048107780516147614,
+      "learning_rate": 0.0001547256928379141,
+      "loss": 0.6998,
+      "step": 1142
+    },
+    {
+      "epoch": 0.64,
+      "grad_norm": 0.04937562346458435,
+      "learning_rate": 0.0001546523477511754,
+      "loss": 0.7552,
+      "step": 1143
+    },
+    {
+      "epoch": 0.64,
+      "grad_norm": 0.05084756389260292,
+      "learning_rate": 0.00015457896071868602,
+      "loss": 0.6517,
+      "step": 1144
+    },
+    {
+      "epoch": 0.64,
+      "grad_norm": 0.04428539052605629,
+      "learning_rate": 0.00015450553179677075,
+      "loss": 0.6143,
+      "step": 1145
+    },
+    {
+      "epoch": 0.64,
+      "grad_norm": 0.050209447741508484,
+      "learning_rate": 0.0001544320610417863,
+      "loss": 0.6805,
+      "step": 1146
+    },
+    {
+      "epoch": 0.64,
+      "grad_norm": 0.04966219514608383,
+      "learning_rate": 0.0001543585485101216,
+      "loss": 0.7239,
+      "step": 1147
+    },
+    {
+      "epoch": 0.64,
+      "grad_norm": 0.0482448972761631,
+      "learning_rate": 0.00015428499425819764,
+      "loss": 0.7492,
+      "step": 1148
+    },
+    {
+      "epoch": 0.64,
+      "grad_norm": 0.05310830846428871,
+      "learning_rate": 0.00015421139834246737,
+      "loss": 0.7678,
+      "step": 1149
+    },
+    {
+      "epoch": 0.64,
+      "grad_norm": 0.04781487584114075,
+      "learning_rate": 0.0001541377608194158,
+      "loss": 0.7321,
+      "step": 1150
+    },
+    {
+      "epoch": 0.64,
+      "grad_norm": 0.050008464604616165,
+      "learning_rate": 0.00015406408174555976,
+      "loss": 0.7224,
+      "step": 1151
+    },
+    {
+      "epoch": 0.64,
+      "grad_norm": 0.047798726707696915,
+      "learning_rate": 0.00015399036117744812,
+      "loss": 0.6562,
+      "step": 1152
+    },
+    {
+      "epoch": 0.64,
+      "grad_norm": 0.04253503307700157,
+      "learning_rate": 0.00015391659917166143,
+      "loss": 0.613,
+      "step": 1153
+    },
+    {
+      "epoch": 0.64,
+      "grad_norm": 0.04759254679083824,
+      "learning_rate": 0.00015384279578481221,
+      "loss": 0.6502,
+      "step": 1154
+    },
+    {
+      "epoch": 0.64,
+      "grad_norm": 0.044068843126297,
+      "learning_rate": 0.00015376895107354464,
+      "loss": 0.6639,
+      "step": 1155
+    },
+    {
+      "epoch": 0.64,
+      "grad_norm": 0.04848353564739227,
+      "learning_rate": 0.00015369506509453458,
+      "loss": 0.6952,
+      "step": 1156
+    },
+    {
+      "epoch": 0.64,
+      "grad_norm": 0.046723756939172745,
+      "learning_rate": 0.00015362113790448967,
+      "loss": 0.5937,
+      "step": 1157
+    },
+    {
+      "epoch": 0.64,
+      "grad_norm": 0.050520990043878555,
+      "learning_rate": 0.0001535471695601491,
+      "loss": 0.7107,
+      "step": 1158
+    },
+    {
+      "epoch": 0.64,
+      "grad_norm": 0.04594268649816513,
+      "learning_rate": 0.00015347316011828373,
+      "loss": 0.6756,
+      "step": 1159
+    },
+    {
+      "epoch": 0.65,
+      "grad_norm": 0.05000408738851547,
+      "learning_rate": 0.00015339910963569584,
+      "loss": 0.6676,
+      "step": 1160
+    },
+    {
+      "epoch": 0.65,
+      "grad_norm": 0.05357789620757103,
+      "learning_rate": 0.00015332501816921928,
+      "loss": 0.7249,
+      "step": 1161
+    },
+    {
+      "epoch": 0.65,
+      "grad_norm": 0.051022421568632126,
+      "learning_rate": 0.00015325088577571939,
+      "loss": 0.7145,
+      "step": 1162
+    },
+    {
+      "epoch": 0.65,
+      "grad_norm": 0.04610150307416916,
+      "learning_rate": 0.00015317671251209285,
+      "loss": 0.6579,
+      "step": 1163
+    },
+    {
+      "epoch": 0.65,
+      "grad_norm": 0.04873840510845184,
+      "learning_rate": 0.00015310249843526776,
+      "loss": 0.7044,
+      "step": 1164
+    },
+    {
+      "epoch": 0.65,
+      "grad_norm": 0.046522945165634155,
+      "learning_rate": 0.00015302824360220353,
+      "loss": 0.6772,
+      "step": 1165
+    },
+    {
+      "epoch": 0.65,
+      "grad_norm": 0.046785563230514526,
+      "learning_rate": 0.0001529539480698908,
+      "loss": 0.6593,
+      "step": 1166
+    },
+    {
+      "epoch": 0.65,
+      "grad_norm": 0.04677551984786987,
+      "learning_rate": 0.00015287961189535155,
+      "loss": 0.7643,
+      "step": 1167
+    },
+    {
+      "epoch": 0.65,
+      "grad_norm": 0.057435374706983566,
+      "learning_rate": 0.00015280523513563885,
+      "loss": 0.7578,
+      "step": 1168
+    },
+    {
+      "epoch": 0.65,
+      "grad_norm": 0.048298779875040054,
+      "learning_rate": 0.000152730817847837,
+      "loss": 0.7108,
+      "step": 1169
+    },
+    {
+      "epoch": 0.65,
+      "grad_norm": 0.04588484391570091,
+      "learning_rate": 0.00015265636008906133,
+      "loss": 0.6706,
+      "step": 1170
+    },
+    {
+      "epoch": 0.65,
+      "grad_norm": 0.05275345966219902,
+      "learning_rate": 0.00015258186191645829,
+      "loss": 0.7065,
+      "step": 1171
+    },
+    {
+      "epoch": 0.65,
+      "grad_norm": 0.045389268547296524,
+      "learning_rate": 0.00015250732338720533,
+      "loss": 0.6161,
+      "step": 1172
+    },
+    {
+      "epoch": 0.65,
+      "grad_norm": 0.051542263478040695,
+      "learning_rate": 0.00015243274455851085,
+      "loss": 0.6771,
+      "step": 1173
+    },
+    {
+      "epoch": 0.65,
+      "grad_norm": 0.04666198417544365,
+      "learning_rate": 0.00015235812548761426,
+      "loss": 0.6068,
+      "step": 1174
+    },
+    {
+      "epoch": 0.65,
+      "grad_norm": 0.04389272257685661,
+      "learning_rate": 0.00015228346623178573,
+      "loss": 0.6682,
+      "step": 1175
+    },
+    {
+      "epoch": 0.65,
+      "grad_norm": 0.044190727174282074,
+      "learning_rate": 0.00015220876684832638,
+      "loss": 0.6145,
+      "step": 1176
+    },
+    {
+      "epoch": 0.65,
+      "grad_norm": 0.049365028738975525,
+      "learning_rate": 0.0001521340273945681,
+      "loss": 0.7827,
+      "step": 1177
+    },
+    {
+      "epoch": 0.66,
+      "grad_norm": 0.04868901148438454,
+      "learning_rate": 0.00015205924792787345,
+      "loss": 0.6776,
+      "step": 1178
+    },
+    {
+      "epoch": 0.66,
+      "grad_norm": 0.04864593222737312,
+      "learning_rate": 0.00015198442850563584,
+      "loss": 0.6541,
+      "step": 1179
+    },
+    {
+      "epoch": 0.66,
+      "grad_norm": 0.046926818788051605,
+      "learning_rate": 0.00015190956918527924,
+      "loss": 0.6965,
+      "step": 1180
+    },
+    {
+      "epoch": 0.66,
+      "grad_norm": 0.044857464730739594,
+      "learning_rate": 0.0001518346700242583,
+      "loss": 0.7031,
+      "step": 1181
+    },
+    {
+      "epoch": 0.66,
+      "grad_norm": 0.05430439114570618,
+      "learning_rate": 0.0001517597310800582,
+      "loss": 0.7288,
+      "step": 1182
+    },
+    {
+      "epoch": 0.66,
+      "grad_norm": 0.05577121302485466,
+      "learning_rate": 0.00015168475241019474,
+      "loss": 0.8773,
+      "step": 1183
+    },
+    {
+      "epoch": 0.66,
+      "grad_norm": 0.046056222170591354,
+      "learning_rate": 0.0001516097340722141,
+      "loss": 0.6732,
+      "step": 1184
+    },
+    {
+      "epoch": 0.66,
+      "grad_norm": 0.05171097069978714,
+      "learning_rate": 0.00015153467612369297,
+      "loss": 0.7971,
+      "step": 1185
+    },
+    {
+      "epoch": 0.66,
+      "grad_norm": 0.04264886677265167,
+      "learning_rate": 0.00015145957862223845,
+      "loss": 0.6446,
+      "step": 1186
+    },
+    {
+      "epoch": 0.66,
+      "grad_norm": 0.04378258436918259,
+      "learning_rate": 0.0001513844416254879,
+      "loss": 0.6927,
+      "step": 1187
+    },
+    {
+      "epoch": 0.66,
+      "grad_norm": 0.04480038583278656,
+      "learning_rate": 0.00015130926519110914,
+      "loss": 0.6879,
+      "step": 1188
+    },
+    {
+      "epoch": 0.66,
+      "grad_norm": 0.04745401442050934,
+      "learning_rate": 0.00015123404937680016,
+      "loss": 0.6483,
+      "step": 1189
+    },
+    {
+      "epoch": 0.66,
+      "grad_norm": 0.050725579261779785,
+      "learning_rate": 0.0001511587942402892,
+      "loss": 0.668,
+      "step": 1190
+    },
+    {
+      "epoch": 0.66,
+      "grad_norm": 0.046093638986349106,
+      "learning_rate": 0.00015108349983933465,
+      "loss": 0.6211,
+      "step": 1191
+    },
+    {
+      "epoch": 0.66,
+      "grad_norm": 0.0595737025141716,
+      "learning_rate": 0.00015100816623172514,
+      "loss": 0.7197,
+      "step": 1192
+    },
+    {
+      "epoch": 0.66,
+      "grad_norm": 0.055872704833745956,
+      "learning_rate": 0.0001509327934752792,
+      "loss": 0.7615,
+      "step": 1193
+    },
+    {
+      "epoch": 0.66,
+      "grad_norm": 0.05438286438584328,
+      "learning_rate": 0.00015085738162784565,
+      "loss": 0.7135,
+      "step": 1194
+    },
+    {
+      "epoch": 0.66,
+      "grad_norm": 0.048681970685720444,
+      "learning_rate": 0.0001507819307473031,
+      "loss": 0.7368,
+      "step": 1195
+    },
+    {
+      "epoch": 0.67,
+      "grad_norm": 0.0487043671309948,
+      "learning_rate": 0.00015070644089156026,
+      "loss": 0.6612,
+      "step": 1196
+    },
+    {
+      "epoch": 0.67,
+      "grad_norm": 0.0529305525124073,
+      "learning_rate": 0.0001506309121185556,
+      "loss": 0.6877,
+      "step": 1197
+    },
+    {
+      "epoch": 0.67,
+      "grad_norm": 0.05354293808341026,
+      "learning_rate": 0.00015055534448625766,
+      "loss": 0.7075,
+      "step": 1198
+    },
+    {
+      "epoch": 0.67,
+      "grad_norm": 0.04844878613948822,
+      "learning_rate": 0.00015047973805266466,
+      "loss": 0.693,
+      "step": 1199
+    },
+    {
+      "epoch": 0.67,
+      "grad_norm": 0.05151020362973213,
+      "learning_rate": 0.0001504040928758046,
+      "loss": 0.6552,
+      "step": 1200
+    },
+    {
+      "epoch": 0.67,
+      "grad_norm": 0.055405277758836746,
+      "learning_rate": 0.00015032840901373532,
+      "loss": 0.7242,
+      "step": 1201
+    },
+    {
+      "epoch": 0.67,
+      "grad_norm": 0.052574001252651215,
+      "learning_rate": 0.00015025268652454421,
+      "loss": 0.7383,
+      "step": 1202
+    },
+    {
+      "epoch": 0.67,
+      "grad_norm": 0.05063820630311966,
+      "learning_rate": 0.0001501769254663485,
+      "loss": 0.6245,
+      "step": 1203
+    },
+    {
+      "epoch": 0.67,
+      "grad_norm": 0.04701600968837738,
+      "learning_rate": 0.00015010112589729482,
+      "loss": 0.6344,
+      "step": 1204
+    },
+    {
+      "epoch": 0.67,
+      "grad_norm": 0.04830322787165642,
+      "learning_rate": 0.00015002528787555944,
+      "loss": 0.6309,
+      "step": 1205
+    },
+    {
+      "epoch": 0.67,
+      "grad_norm": 0.04776900261640549,
+      "learning_rate": 0.00014994941145934814,
+      "loss": 0.7883,
+      "step": 1206
+    },
+    {
+      "epoch": 0.67,
+      "grad_norm": 0.048602622002363205,
+      "learning_rate": 0.00014987349670689623,
+      "loss": 0.7106,
+      "step": 1207
+    },
+    {
+      "epoch": 0.67,
+      "grad_norm": 0.05431029573082924,
+      "learning_rate": 0.00014979754367646834,
+      "loss": 0.7456,
+      "step": 1208
+    },
+    {
+      "epoch": 0.67,
+      "grad_norm": 0.04521273076534271,
+      "learning_rate": 0.00014972155242635852,
+      "loss": 0.5661,
+      "step": 1209
+    },
+    {
+      "epoch": 0.67,
+      "grad_norm": 0.04839641973376274,
+      "learning_rate": 0.0001496455230148902,
+      "loss": 0.6362,
+      "step": 1210
+    },
+    {
+      "epoch": 0.67,
+      "grad_norm": 0.05457577481865883,
+      "learning_rate": 0.00014956945550041595,
+      "loss": 0.792,
+      "step": 1211
+    },
+    {
+      "epoch": 0.67,
+      "grad_norm": 0.051022760570049286,
+      "learning_rate": 0.0001494933499413178,
+      "loss": 0.7257,
+      "step": 1212
+    },
+    {
+      "epoch": 0.67,
+      "grad_norm": 0.05146949738264084,
+      "learning_rate": 0.00014941720639600686,
+      "loss": 0.7664,
+      "step": 1213
+    },
+    {
+      "epoch": 0.68,
+      "grad_norm": 0.04937746375799179,
+      "learning_rate": 0.00014934102492292336,
+      "loss": 0.7638,
+      "step": 1214
+    },
+    {
+      "epoch": 0.68,
+      "grad_norm": 0.05288033187389374,
+      "learning_rate": 0.0001492648055805367,
+      "loss": 0.7008,
+      "step": 1215
+    },
+    {
+      "epoch": 0.68,
+      "grad_norm": 0.04645581543445587,
+      "learning_rate": 0.00014918854842734533,
+      "loss": 0.659,
+      "step": 1216
+    },
+    {
+      "epoch": 0.68,
+      "grad_norm": 0.049302008002996445,
+      "learning_rate": 0.0001491122535218767,
+      "loss": 0.6522,
+      "step": 1217
+    },
+    {
+      "epoch": 0.68,
+      "grad_norm": 0.047872643917798996,
+      "learning_rate": 0.00014903592092268728,
+      "loss": 0.7293,
+      "step": 1218
+    },
+    {
+      "epoch": 0.68,
+      "grad_norm": 0.046734388917684555,
+      "learning_rate": 0.00014895955068836242,
+      "loss": 0.5885,
+      "step": 1219
+    },
+    {
+      "epoch": 0.68,
+      "grad_norm": 0.05098286643624306,
+      "learning_rate": 0.0001488831428775164,
+      "loss": 0.6703,
+      "step": 1220
+    },
+    {
+      "epoch": 0.68,
+      "grad_norm": 0.05204523354768753,
+      "learning_rate": 0.00014880669754879227,
+      "loss": 0.7022,
+      "step": 1221
+    },
+    {
+      "epoch": 0.68,
+      "grad_norm": 0.049951422959566116,
+      "learning_rate": 0.00014873021476086203,
+      "loss": 0.6648,
+      "step": 1222
+    },
+    {
+      "epoch": 0.68,
+      "grad_norm": 0.04843631759285927,
+      "learning_rate": 0.00014865369457242621,
+      "loss": 0.6847,
+      "step": 1223
+    },
+    {
+      "epoch": 0.68,
+      "grad_norm": 0.05291084572672844,
+      "learning_rate": 0.00014857713704221422,
+      "loss": 0.8186,
+      "step": 1224
+    },
+    {
+      "epoch": 0.68,
+      "grad_norm": 0.05095936357975006,
+      "learning_rate": 0.00014850054222898402,
+      "loss": 0.705,
+      "step": 1225
+    },
+    {
+      "epoch": 0.68,
+      "grad_norm": 0.05311046168208122,
+      "learning_rate": 0.00014842391019152226,
+      "loss": 0.6515,
+      "step": 1226
+    },
+    {
+      "epoch": 0.68,
+      "grad_norm": 0.05315253511071205,
+      "learning_rate": 0.00014834724098864417,
+      "loss": 0.6898,
+      "step": 1227
+    },
+    {
+      "epoch": 0.68,
+      "grad_norm": 0.048906102776527405,
+      "learning_rate": 0.0001482705346791934,
+      "loss": 0.7955,
+      "step": 1228
+    },
+    {
+      "epoch": 0.68,
+      "grad_norm": 0.0493389368057251,
+      "learning_rate": 0.00014819379132204218,
+      "loss": 0.7394,
+      "step": 1229
+    },
+    {
+      "epoch": 0.68,
+      "grad_norm": 0.04702428728342056,
+      "learning_rate": 0.00014811701097609112,
+      "loss": 0.5547,
+      "step": 1230
+    },
+    {
+      "epoch": 0.68,
+      "grad_norm": 0.04860011860728264,
+      "learning_rate": 0.00014804019370026926,
+      "loss": 0.7455,
+      "step": 1231
+    },
+    {
+      "epoch": 0.69,
+      "grad_norm": 0.05252429470419884,
+      "learning_rate": 0.00014796333955353395,
+      "loss": 0.7699,
+      "step": 1232
+    },
+    {
+      "epoch": 0.69,
+      "grad_norm": 0.04439985752105713,
+      "learning_rate": 0.0001478864485948709,
+      "loss": 0.6318,
+      "step": 1233
+    },
+    {
+      "epoch": 0.69,
+      "grad_norm": 0.051365215331315994,
+      "learning_rate": 0.00014780952088329396,
+      "loss": 0.748,
+      "step": 1234
+    },
+    {
+      "epoch": 0.69,
+      "grad_norm": 0.050039131194353104,
+      "learning_rate": 0.00014773255647784525,
+      "loss": 0.6782,
+      "step": 1235
+    },
+    {
+      "epoch": 0.69,
+      "grad_norm": 0.04989631101489067,
+      "learning_rate": 0.0001476555554375951,
+      "loss": 0.6101,
+      "step": 1236
+    },
+    {
+      "epoch": 0.69,
+      "grad_norm": 0.05142849683761597,
+      "learning_rate": 0.0001475785178216419,
+      "loss": 0.6938,
+      "step": 1237
+    },
+    {
+      "epoch": 0.69,
+      "grad_norm": 0.04728643596172333,
+      "learning_rate": 0.00014750144368911209,
+      "loss": 0.6917,
+      "step": 1238
+    },
+    {
+      "epoch": 0.69,
+      "grad_norm": 0.0472208671271801,
+      "learning_rate": 0.0001474243330991602,
+      "loss": 0.6225,
+      "step": 1239
+    },
+    {
+      "epoch": 0.69,
+      "grad_norm": 0.05022706091403961,
+      "learning_rate": 0.00014734718611096877,
+      "loss": 0.6543,
+      "step": 1240
+    },
+    {
+      "epoch": 0.69,
+      "grad_norm": 0.05018246918916702,
+      "learning_rate": 0.00014727000278374808,
+      "loss": 0.7208,
+      "step": 1241
+    },
+    {
+      "epoch": 0.69,
+      "grad_norm": 0.04962945356965065,
+      "learning_rate": 0.00014719278317673655,
+      "loss": 0.6386,
+      "step": 1242
+    },
+    {
+      "epoch": 0.69,
+      "grad_norm": 0.04697633534669876,
+      "learning_rate": 0.0001471155273492003,
+      "loss": 0.6764,
+      "step": 1243
+    },
+    {
+      "epoch": 0.69,
+      "grad_norm": 0.05113474652171135,
+      "learning_rate": 0.00014703823536043323,
+      "loss": 0.7663,
+      "step": 1244
+    },
+    {
+      "epoch": 0.69,
+      "grad_norm": 0.05091071128845215,
+      "learning_rate": 0.00014696090726975714,
+      "loss": 0.7107,
+      "step": 1245
+    },
+    {
+      "epoch": 0.69,
+      "grad_norm": 0.05539362505078316,
+      "learning_rate": 0.0001468835431365214,
+      "loss": 0.6792,
+      "step": 1246
+    },
+    {
+      "epoch": 0.69,
+      "grad_norm": 0.05495639145374298,
+      "learning_rate": 0.0001468061430201031,
+      "loss": 0.7616,
+      "step": 1247
+    },
+    {
+      "epoch": 0.69,
+      "grad_norm": 0.05079134553670883,
+      "learning_rate": 0.00014672870697990685,
+      "loss": 0.6765,
+      "step": 1248
+    },
+    {
+      "epoch": 0.69,
+      "grad_norm": 0.048591259866952896,
+      "learning_rate": 0.00014665123507536505,
+      "loss": 0.6775,
+      "step": 1249
+    },
+    {
+      "epoch": 0.7,
+      "grad_norm": 0.04820895567536354,
+      "learning_rate": 0.00014657372736593739,
+      "loss": 0.6892,
+      "step": 1250
+    },
+    {
+      "epoch": 0.7,
+      "grad_norm": 0.0469762422144413,
+      "learning_rate": 0.00014649618391111117,
+      "loss": 0.6423,
+      "step": 1251
+    },
+    {
+      "epoch": 0.7,
+      "grad_norm": 0.048589132726192474,
+      "learning_rate": 0.00014641860477040112,
+      "loss": 0.6854,
+      "step": 1252
+    },
+    {
+      "epoch": 0.7,
+      "grad_norm": 0.05065524950623512,
+      "learning_rate": 0.0001463409900033493,
+      "loss": 0.6663,
+      "step": 1253
+    },
+    {
+      "epoch": 0.7,
+      "grad_norm": 0.049449969083070755,
+      "learning_rate": 0.0001462633396695252,
+      "loss": 0.6662,
+      "step": 1254
+    },
+    {
+      "epoch": 0.7,
+      "grad_norm": 0.052549589425325394,
+      "learning_rate": 0.00014618565382852548,
+      "loss": 0.7664,
+      "step": 1255
+    },
+    {
+      "epoch": 0.7,
+      "grad_norm": 0.05117465555667877,
+      "learning_rate": 0.00014610793253997423,
+      "loss": 0.7495,
+      "step": 1256
+    },
+    {
+      "epoch": 0.7,
+      "grad_norm": 0.054655950516462326,
+      "learning_rate": 0.0001460301758635225,
+      "loss": 0.6196,
+      "step": 1257
+    },
+    {
+      "epoch": 0.7,
+      "grad_norm": 0.049078892916440964,
+      "learning_rate": 0.0001459523838588488,
+      "loss": 0.5712,
+      "step": 1258
+    },
+    {
+      "epoch": 0.7,
+      "grad_norm": 0.05013487488031387,
+      "learning_rate": 0.00014587455658565847,
+      "loss": 0.6546,
+      "step": 1259
+    },
+    {
+      "epoch": 0.7,
+      "grad_norm": 0.05131218209862709,
+      "learning_rate": 0.00014579669410368413,
+      "loss": 0.6762,
+      "step": 1260
+    },
+    {
+      "epoch": 0.7,
+      "grad_norm": 0.05478007718920708,
+      "learning_rate": 0.00014571879647268528,
+      "loss": 0.6899,
+      "step": 1261
+    },
+    {
+      "epoch": 0.7,
+      "grad_norm": 0.04590092971920967,
+      "learning_rate": 0.00014564086375244855,
+      "loss": 0.6682,
+      "step": 1262
+    },
+    {
+      "epoch": 0.7,
+      "grad_norm": 0.055836863815784454,
+      "learning_rate": 0.00014556289600278733,
+      "loss": 0.8131,
+      "step": 1263
+    },
+    {
+      "epoch": 0.7,
+      "grad_norm": 0.05465158447623253,
+      "learning_rate": 0.00014548489328354195,
+      "loss": 0.7749,
+      "step": 1264
+    },
+    {
+      "epoch": 0.7,
+      "grad_norm": 0.04707031324505806,
+      "learning_rate": 0.00014540685565457968,
+      "loss": 0.5979,
+      "step": 1265
+    },
+    {
+      "epoch": 0.7,
+      "grad_norm": 0.048351485282182693,
+      "learning_rate": 0.00014532878317579444,
+      "loss": 0.7093,
+      "step": 1266
+    },
+    {
+      "epoch": 0.7,
+      "grad_norm": 0.05322815105319023,
+      "learning_rate": 0.000145250675907107,
+      "loss": 0.7281,
+      "step": 1267
+    },
+    {
+      "epoch": 0.71,
+      "grad_norm": 0.04762796312570572,
+      "learning_rate": 0.0001451725339084648,
+      "loss": 0.6659,
+      "step": 1268
+    },
+    {
+      "epoch": 0.71,
+      "grad_norm": 0.047569163143634796,
+      "learning_rate": 0.0001450943572398419,
+      "loss": 0.6894,
+      "step": 1269
+    },
+    {
+      "epoch": 0.71,
+      "grad_norm": 0.04456879943609238,
+      "learning_rate": 0.00014501614596123898,
+      "loss": 0.6958,
+      "step": 1270
+    },
+    {
+      "epoch": 0.71,
+      "grad_norm": 0.050228893756866455,
+      "learning_rate": 0.00014493790013268338,
+      "loss": 0.7011,
+      "step": 1271
+    },
+    {
+      "epoch": 0.71,
+      "grad_norm": 0.056763775646686554,
+      "learning_rate": 0.0001448596198142288,
+      "loss": 0.8265,
+      "step": 1272
+    },
+    {
+      "epoch": 0.71,
+      "grad_norm": 0.04859040677547455,
+      "learning_rate": 0.00014478130506595556,
+      "loss": 0.7295,
+      "step": 1273
+    },
+    {
+      "epoch": 0.71,
+      "grad_norm": 0.054411161690950394,
+      "learning_rate": 0.0001447029559479703,
+      "loss": 0.7045,
+      "step": 1274
+    },
+    {
+      "epoch": 0.71,
+      "grad_norm": 0.047824714332818985,
+      "learning_rate": 0.00014462457252040607,
+      "loss": 0.7665,
+      "step": 1275
+    },
+    {
+      "epoch": 0.71,
+      "grad_norm": 0.04883917048573494,
+      "learning_rate": 0.00014454615484342222,
+      "loss": 0.7283,
+      "step": 1276
+    },
+    {
+      "epoch": 0.71,
+      "grad_norm": 0.04813641682267189,
+      "learning_rate": 0.00014446770297720448,
+      "loss": 0.688,
+      "step": 1277
+    },
+    {
+      "epoch": 0.71,
+      "grad_norm": 0.049928583204746246,
+      "learning_rate": 0.00014438921698196476,
+      "loss": 0.7347,
+      "step": 1278
+    },
+    {
+      "epoch": 0.71,
+      "grad_norm": 0.05147488787770271,
+      "learning_rate": 0.00014431069691794114,
+      "loss": 0.7492,
+      "step": 1279
+    },
+    {
+      "epoch": 0.71,
+      "grad_norm": 0.04984632879495621,
+      "learning_rate": 0.0001442321428453979,
+      "loss": 0.744,
+      "step": 1280
+    },
+    {
+      "epoch": 0.71,
+      "grad_norm": 0.05544523149728775,
+      "learning_rate": 0.00014415355482462534,
+      "loss": 0.7598,
+      "step": 1281
+    },
+    {
+      "epoch": 0.71,
+      "grad_norm": 0.05009017884731293,
+      "learning_rate": 0.00014407493291593995,
+      "loss": 0.6616,
+      "step": 1282
+    },
+    {
+      "epoch": 0.71,
+      "grad_norm": 0.050635501742362976,
+      "learning_rate": 0.00014399627717968412,
+      "loss": 0.6241,
+      "step": 1283
+    },
+    {
+      "epoch": 0.71,
+      "grad_norm": 0.054587721824645996,
+      "learning_rate": 0.0001439175876762262,
+      "loss": 0.6985,
+      "step": 1284
+    },
+    {
+      "epoch": 0.71,
+      "grad_norm": 0.05211169272661209,
+      "learning_rate": 0.00014383886446596057,
+      "loss": 0.6875,
+      "step": 1285
+    },
+    {
+      "epoch": 0.72,
+      "grad_norm": 0.05252880975604057,
+      "learning_rate": 0.00014376010760930728,
+      "loss": 0.6947,
+      "step": 1286
+    },
+    {
+      "epoch": 0.72,
+      "grad_norm": 0.04912309721112251,
+      "learning_rate": 0.00014368131716671246,
+      "loss": 0.6376,
+      "step": 1287
+    },
+    {
+      "epoch": 0.72,
+      "grad_norm": 0.05693266540765762,
+      "learning_rate": 0.00014360249319864775,
+      "loss": 0.7267,
+      "step": 1288
+    },
+    {
+      "epoch": 0.72,
+      "grad_norm": 0.05367673560976982,
+      "learning_rate": 0.00014352363576561073,
+      "loss": 0.714,
+      "step": 1289
+    },
+    {
+      "epoch": 0.72,
+      "grad_norm": 0.04969675466418266,
+      "learning_rate": 0.00014344474492812461,
+      "loss": 0.6023,
+      "step": 1290
+    },
+    {
+      "epoch": 0.72,
+      "grad_norm": 0.04907987639307976,
+      "learning_rate": 0.00014336582074673813,
+      "loss": 0.7009,
+      "step": 1291
+    },
+    {
+      "epoch": 0.72,
+      "grad_norm": 0.04774461314082146,
+      "learning_rate": 0.00014328686328202582,
+      "loss": 0.6393,
+      "step": 1292
+    },
+    {
+      "epoch": 0.72,
+      "grad_norm": 0.047509051859378815,
+      "learning_rate": 0.00014320787259458753,
+      "loss": 0.6444,
+      "step": 1293
+    },
+    {
+      "epoch": 0.72,
+      "grad_norm": 0.057917118072509766,
+      "learning_rate": 0.00014312884874504876,
+      "loss": 0.8577,
+      "step": 1294
+    },
+    {
+      "epoch": 0.72,
+      "grad_norm": 0.04852381348609924,
+      "learning_rate": 0.0001430497917940604,
+      "loss": 0.6916,
+      "step": 1295
+    },
+    {
+      "epoch": 0.72,
+      "grad_norm": 0.04885304719209671,
+      "learning_rate": 0.00014297070180229883,
+      "loss": 0.7174,
+      "step": 1296
+    },
+    {
+      "epoch": 0.72,
+      "grad_norm": 0.052427150309085846,
+      "learning_rate": 0.00014289157883046568,
+      "loss": 0.8375,
+      "step": 1297
+    },
+    {
+      "epoch": 0.72,
+      "grad_norm": 0.04951222613453865,
+      "learning_rate": 0.0001428124229392879,
+      "loss": 0.7188,
+      "step": 1298
+    },
+    {
+      "epoch": 0.72,
+      "grad_norm": 0.049055252224206924,
+      "learning_rate": 0.0001427332341895178,
+      "loss": 0.6977,
+      "step": 1299
+    },
+    {
+      "epoch": 0.72,
+      "grad_norm": 0.047418877482414246,
+      "learning_rate": 0.00014265401264193286,
+      "loss": 0.5577,
+      "step": 1300
+    },
+    {
+      "epoch": 0.72,
+      "grad_norm": 0.051448144018650055,
+      "learning_rate": 0.00014257475835733572,
+      "loss": 0.6462,
+      "step": 1301
+    },
+    {
+      "epoch": 0.72,
+      "grad_norm": 0.053073249757289886,
+      "learning_rate": 0.00014249547139655409,
+      "loss": 0.6129,
+      "step": 1302
+    },
+    {
+      "epoch": 0.72,
+      "grad_norm": 0.05240590125322342,
+      "learning_rate": 0.00014241615182044088,
+      "loss": 0.7354,
+      "step": 1303
+    },
+    {
+      "epoch": 0.73,
+      "grad_norm": 0.04916330799460411,
+      "learning_rate": 0.00014233679968987393,
+      "loss": 0.6193,
+      "step": 1304
+    },
+    {
+      "epoch": 0.73,
+      "grad_norm": 0.046741846948862076,
+      "learning_rate": 0.00014225741506575617,
+      "loss": 0.5996,
+      "step": 1305
+    },
+    {
+      "epoch": 0.73,
+      "grad_norm": 0.050647541880607605,
+      "learning_rate": 0.00014217799800901533,
+      "loss": 0.6619,
+      "step": 1306
+    },
+    {
+      "epoch": 0.73,
+      "grad_norm": 0.052585843950510025,
+      "learning_rate": 0.00014209854858060415,
+      "loss": 0.6119,
+      "step": 1307
+    },
+    {
+      "epoch": 0.73,
+      "grad_norm": 0.05018146336078644,
+      "learning_rate": 0.0001420190668415002,
+      "loss": 0.6795,
+      "step": 1308
+    },
+    {
+      "epoch": 0.73,
+      "grad_norm": 0.052429914474487305,
+      "learning_rate": 0.00014193955285270575,
+      "loss": 0.6865,
+      "step": 1309
+    },
+    {
+      "epoch": 0.73,
+      "grad_norm": 0.051422689110040665,
+      "learning_rate": 0.00014186000667524795,
+      "loss": 0.7339,
+      "step": 1310
+    },
+    {
+      "epoch": 0.73,
+      "grad_norm": 0.05736821889877319,
+      "learning_rate": 0.00014178042837017856,
+      "loss": 0.7109,
+      "step": 1311
+    },
+    {
+      "epoch": 0.73,
+      "grad_norm": 0.04986963048577309,
+      "learning_rate": 0.0001417008179985741,
+      "loss": 0.6964,
+      "step": 1312
+    },
+    {
+      "epoch": 0.73,
+      "grad_norm": 0.05136452242732048,
+      "learning_rate": 0.00014162117562153558,
+      "loss": 0.5924,
+      "step": 1313
+    },
+    {
+      "epoch": 0.73,
+      "grad_norm": 0.05401051789522171,
+      "learning_rate": 0.00014154150130018866,
+      "loss": 0.756,
+      "step": 1314
+    },
+    {
+      "epoch": 0.73,
+      "grad_norm": 0.05107508972287178,
+      "learning_rate": 0.00014146179509568344,
+      "loss": 0.6913,
+      "step": 1315
+    },
+    {
+      "epoch": 0.73,
+      "grad_norm": 0.048386212438344955,
+      "learning_rate": 0.0001413820570691946,
+      "loss": 0.6138,
+      "step": 1316
+    },
+    {
+      "epoch": 0.73,
+      "grad_norm": 0.056987110525369644,
+      "learning_rate": 0.00014130228728192118,
+      "loss": 0.6351,
+      "step": 1317
+    },
+    {
+      "epoch": 0.73,
+      "grad_norm": 0.04835304245352745,
+      "learning_rate": 0.00014122248579508657,
+      "loss": 0.6808,
+      "step": 1318
+    },
+    {
+      "epoch": 0.73,
+      "grad_norm": 0.0571272112429142,
+      "learning_rate": 0.00014114265266993846,
+      "loss": 0.6881,
+      "step": 1319
+    },
+    {
+      "epoch": 0.73,
+      "grad_norm": 0.058113086968660355,
+      "learning_rate": 0.00014106278796774903,
+      "loss": 0.7849,
+      "step": 1320
+    },
+    {
+      "epoch": 0.73,
+      "grad_norm": 0.05638326704502106,
+      "learning_rate": 0.00014098289174981443,
+      "loss": 0.6941,
+      "step": 1321
+    },
+    {
+      "epoch": 0.74,
+      "grad_norm": 0.05447527393698692,
+      "learning_rate": 0.00014090296407745517,
+      "loss": 0.7279,
+      "step": 1322
+    },
+    {
+      "epoch": 0.74,
+      "grad_norm": 0.05510277301073074,
+      "learning_rate": 0.0001408230050120158,
+      "loss": 0.7933,
+      "step": 1323
+    },
+    {
+      "epoch": 0.74,
+      "grad_norm": 0.059467945247888565,
+      "learning_rate": 0.00014074301461486504,
+      "loss": 0.7845,
+      "step": 1324
+    },
+    {
+      "epoch": 0.74,
+      "grad_norm": 0.060325901955366135,
+      "learning_rate": 0.00014066299294739567,
+      "loss": 0.7254,
+      "step": 1325
+    },
+    {
+      "epoch": 0.74,
+      "grad_norm": 0.05181081220507622,
+      "learning_rate": 0.00014058294007102432,
+      "loss": 0.6383,
+      "step": 1326
+    },
+    {
+      "epoch": 0.74,
+      "grad_norm": 0.050908952951431274,
+      "learning_rate": 0.00014050285604719177,
+      "loss": 0.7441,
+      "step": 1327
+    },
+    {
+      "epoch": 0.74,
+      "grad_norm": 0.049533553421497345,
+      "learning_rate": 0.0001404227409373626,
+      "loss": 0.6214,
+      "step": 1328
+    },
+    {
+      "epoch": 0.74,
+      "grad_norm": 0.05213542655110359,
+      "learning_rate": 0.00014034259480302523,
+      "loss": 0.6518,
+      "step": 1329
+    },
+    {
+      "epoch": 0.74,
+      "grad_norm": 0.05297151952981949,
+      "learning_rate": 0.00014026241770569197,
+      "loss": 0.7337,
+      "step": 1330
+    },
+    {
+      "epoch": 0.74,
+      "grad_norm": 0.04704327508807182,
+      "learning_rate": 0.00014018220970689883,
+      "loss": 0.619,
+      "step": 1331
+    },
+    {
+      "epoch": 0.74,
+      "grad_norm": 0.054115235805511475,
+      "learning_rate": 0.00014010197086820552,
+      "loss": 0.7122,
+      "step": 1332
+    },
+    {
+      "epoch": 0.74,
+      "grad_norm": 0.050988852977752686,
+      "learning_rate": 0.00014002170125119553,
+      "loss": 0.7545,
+      "step": 1333
+    },
+    {
+      "epoch": 0.74,
+      "grad_norm": 0.049325957894325256,
+      "learning_rate": 0.00013994140091747587,
+      "loss": 0.6834,
+      "step": 1334
+    },
+    {
+      "epoch": 0.74,
+      "grad_norm": 0.06573436409235,
+      "learning_rate": 0.00013986106992867713,
+      "loss": 0.7652,
+      "step": 1335
+    },
+    {
+      "epoch": 0.74,
+      "grad_norm": 0.050615664571523666,
+      "learning_rate": 0.00013978070834645348,
+      "loss": 0.648,
+      "step": 1336
+    },
+    {
+      "epoch": 0.74,
+      "grad_norm": 0.050100456923246384,
+      "learning_rate": 0.0001397003162324825,
+      "loss": 0.6443,
+      "step": 1337
+    },
+    {
+      "epoch": 0.74,
+      "grad_norm": 0.04777170345187187,
+      "learning_rate": 0.00013961989364846532,
+      "loss": 0.6743,
+      "step": 1338
+    },
+    {
+      "epoch": 0.74,
+      "grad_norm": 0.05937036871910095,
+      "learning_rate": 0.00013953944065612633,
+      "loss": 0.7196,
+      "step": 1339
+    },
+    {
+      "epoch": 0.75,
+      "grad_norm": 0.05465259775519371,
+      "learning_rate": 0.0001394589573172133,
+      "loss": 0.7038,
+      "step": 1340
+    },
+    {
+      "epoch": 0.75,
+      "grad_norm": 0.04839110001921654,
+      "learning_rate": 0.00013937844369349734,
+      "loss": 0.7035,
+      "step": 1341
+    },
+    {
+      "epoch": 0.75,
+      "grad_norm": 0.05537423491477966,
+      "learning_rate": 0.00013929789984677278,
+      "loss": 0.7386,
+      "step": 1342
+    },
+    {
+      "epoch": 0.75,
+      "grad_norm": 0.051877472549676895,
+      "learning_rate": 0.00013921732583885705,
+      "loss": 0.6811,
+      "step": 1343
+    },
+    {
+      "epoch": 0.75,
+      "grad_norm": 0.052817557007074356,
+      "learning_rate": 0.00013913672173159088,
+      "loss": 0.6852,
+      "step": 1344
+    },
+    {
+      "epoch": 0.75,
+      "grad_norm": 0.0512346476316452,
+      "learning_rate": 0.000139056087586838,
+      "loss": 0.6145,
+      "step": 1345
+    },
+    {
+      "epoch": 0.75,
+      "grad_norm": 0.05505014955997467,
+      "learning_rate": 0.00013897542346648524,
+      "loss": 0.8609,
+      "step": 1346
+    },
+    {
+      "epoch": 0.75,
+      "grad_norm": 0.0500144399702549,
+      "learning_rate": 0.00013889472943244243,
+      "loss": 0.6782,
+      "step": 1347
+    },
+    {
+      "epoch": 0.75,
+      "grad_norm": 0.0514250211417675,
+      "learning_rate": 0.0001388140055466423,
+      "loss": 0.7126,
+      "step": 1348
+    },
+    {
+      "epoch": 0.75,
+      "grad_norm": 0.049233101308345795,
+      "learning_rate": 0.00013873325187104056,
+      "loss": 0.6516,
+      "step": 1349
+    },
+    {
+      "epoch": 0.75,
+      "grad_norm": 0.04899629205465317,
+      "learning_rate": 0.00013865246846761581,
+      "loss": 0.617,
+      "step": 1350
+    },
+    {
+      "epoch": 0.75,
+      "grad_norm": 0.05683201178908348,
+      "learning_rate": 0.0001385716553983694,
+      "loss": 0.7033,
+      "step": 1351
+    },
+    {
+      "epoch": 0.75,
+      "grad_norm": 0.05018038675189018,
+      "learning_rate": 0.00013849081272532544,
+      "loss": 0.6809,
+      "step": 1352
+    },
+    {
+      "epoch": 0.75,
+      "grad_norm": 0.05704934149980545,
+      "learning_rate": 0.00013840994051053085,
+      "loss": 0.8067,
+      "step": 1353
+    },
+    {
+      "epoch": 0.75,
+      "grad_norm": 0.050851497799158096,
+      "learning_rate": 0.00013832903881605508,
+      "loss": 0.6693,
+      "step": 1354
+    },
+    {
+      "epoch": 0.75,
+      "grad_norm": 0.051365941762924194,
+      "learning_rate": 0.00013824810770399036,
+      "loss": 0.6739,
+      "step": 1355
+    },
+    {
+      "epoch": 0.75,
+      "grad_norm": 0.058710698038339615,
+      "learning_rate": 0.0001381671472364514,
+      "loss": 0.6647,
+      "step": 1356
+    },
+    {
+      "epoch": 0.75,
+      "grad_norm": 0.05513182282447815,
+      "learning_rate": 0.00013808615747557549,
+      "loss": 0.6949,
+      "step": 1357
+    },
+    {
+      "epoch": 0.76,
+      "grad_norm": 0.05616655945777893,
+      "learning_rate": 0.0001380051384835223,
+      "loss": 0.7774,
+      "step": 1358
+    },
+    {
+      "epoch": 0.76,
+      "grad_norm": 0.05610072240233421,
+      "learning_rate": 0.0001379240903224741,
+      "loss": 0.7263,
+      "step": 1359
+    },
+    {
+      "epoch": 0.76,
+      "grad_norm": 0.05135612562298775,
+      "learning_rate": 0.00013784301305463549,
+      "loss": 0.6889,
+      "step": 1360
+    },
+    {
+      "epoch": 0.76,
+      "grad_norm": 0.05268790200352669,
+      "learning_rate": 0.00013776190674223327,
+      "loss": 0.7119,
+      "step": 1361
+    },
+    {
+      "epoch": 0.76,
+      "grad_norm": 0.04713091999292374,
+      "learning_rate": 0.00013768077144751673,
+      "loss": 0.6209,
+      "step": 1362
+    },
+    {
+      "epoch": 0.76,
+      "grad_norm": 0.049119122326374054,
+      "learning_rate": 0.00013759960723275732,
+      "loss": 0.673,
+      "step": 1363
+    },
+    {
+      "epoch": 0.76,
+      "grad_norm": 0.05238724872469902,
+      "learning_rate": 0.00013751841416024865,
+      "loss": 0.7073,
+      "step": 1364
+    },
+    {
+      "epoch": 0.76,
+      "grad_norm": 0.04365584999322891,
+      "learning_rate": 0.0001374371922923065,
+      "loss": 0.5371,
+      "step": 1365
+    },
+    {
+      "epoch": 0.76,
+      "grad_norm": 0.05708995461463928,
+      "learning_rate": 0.0001373559416912688,
+      "loss": 0.6894,
+      "step": 1366
+    },
+    {
+      "epoch": 0.76,
+      "grad_norm": 0.04669976979494095,
+      "learning_rate": 0.00013727466241949545,
+      "loss": 0.7083,
+      "step": 1367
+    },
+    {
+      "epoch": 0.76,
+      "grad_norm": 0.05093741416931152,
+      "learning_rate": 0.00013719335453936846,
+      "loss": 0.6603,
+      "step": 1368
+    },
+    {
+      "epoch": 0.76,
+      "grad_norm": 0.04979169741272926,
+      "learning_rate": 0.0001371120181132917,
+      "loss": 0.6269,
+      "step": 1369
+    },
+    {
+      "epoch": 0.76,
+      "grad_norm": 0.05657560005784035,
+      "learning_rate": 0.000137030653203691,
+      "loss": 0.8572,
+      "step": 1370
+    },
+    {
+      "epoch": 0.76,
+      "grad_norm": 0.05758770555257797,
+      "learning_rate": 0.00013694925987301404,
+      "loss": 0.7561,
+      "step": 1371
+    },
+    {
+      "epoch": 0.76,
+      "grad_norm": 0.048717837780714035,
+      "learning_rate": 0.00013686783818373028,
+      "loss": 0.7044,
+      "step": 1372
+    },
+    {
+      "epoch": 0.76,
+      "grad_norm": 0.049861840903759,
+      "learning_rate": 0.00013678638819833103,
+      "loss": 0.6199,
+      "step": 1373
+    },
+    {
+      "epoch": 0.76,
+      "grad_norm": 0.050144508481025696,
+      "learning_rate": 0.00013670490997932922,
+      "loss": 0.6802,
+      "step": 1374
+    },
+    {
+      "epoch": 0.76,
+      "grad_norm": 0.05109990015625954,
+      "learning_rate": 0.0001366234035892595,
+      "loss": 0.6872,
+      "step": 1375
+    },
+    {
+      "epoch": 0.77,
+      "grad_norm": 0.05122920870780945,
+      "learning_rate": 0.00013654186909067817,
+      "loss": 0.6246,
+      "step": 1376
+    },
+    {
+      "epoch": 0.77,
+      "grad_norm": 0.05392398312687874,
+      "learning_rate": 0.00013646030654616302,
+      "loss": 0.6679,
+      "step": 1377
+    },
+    {
+      "epoch": 0.77,
+      "grad_norm": 0.051943257451057434,
+      "learning_rate": 0.0001363787160183134,
+      "loss": 0.6608,
+      "step": 1378
+    },
+    {
+      "epoch": 0.77,
+      "grad_norm": 0.05808079242706299,
+      "learning_rate": 0.00013629709756975023,
+      "loss": 0.6256,
+      "step": 1379
+    },
+    {
+      "epoch": 0.77,
+      "grad_norm": 0.05690327659249306,
+      "learning_rate": 0.0001362154512631157,
+      "loss": 0.6313,
+      "step": 1380
+    },
+    {
+      "epoch": 0.77,
+      "grad_norm": 0.0537363737821579,
+      "learning_rate": 0.0001361337771610735,
+      "loss": 0.6827,
+      "step": 1381
+    },
+    {
+      "epoch": 0.77,
+      "grad_norm": 0.05171459913253784,
+      "learning_rate": 0.00013605207532630864,
+      "loss": 0.6683,
+      "step": 1382
+    },
+    {
+      "epoch": 0.77,
+      "grad_norm": 0.051463544368743896,
+      "learning_rate": 0.00013597034582152733,
+      "loss": 0.6894,
+      "step": 1383
+    },
+    {
+      "epoch": 0.77,
+      "grad_norm": 0.049201156944036484,
+      "learning_rate": 0.0001358885887094571,
+      "loss": 0.6486,
+      "step": 1384
+    },
+    {
+      "epoch": 0.77,
+      "grad_norm": 0.052608225494623184,
+      "learning_rate": 0.00013580680405284664,
+      "loss": 0.7302,
+      "step": 1385
+    },
+    {
+      "epoch": 0.77,
+      "grad_norm": 0.05633961036801338,
+      "learning_rate": 0.00013572499191446578,
+      "loss": 0.7057,
+      "step": 1386
+    },
+    {
+      "epoch": 0.77,
+      "grad_norm": 0.05480135232210159,
+      "learning_rate": 0.00013564315235710546,
+      "loss": 0.6837,
+      "step": 1387
+    },
+    {
+      "epoch": 0.77,
+      "grad_norm": 0.04823005944490433,
+      "learning_rate": 0.00013556128544357763,
+      "loss": 0.6007,
+      "step": 1388
+    },
+    {
+      "epoch": 0.77,
+      "grad_norm": 0.05478844791650772,
+      "learning_rate": 0.0001354793912367153,
+      "loss": 0.6724,
+      "step": 1389
+    },
+    {
+      "epoch": 0.77,
+      "grad_norm": 0.04951076954603195,
+      "learning_rate": 0.00013539746979937233,
+      "loss": 0.6648,
+      "step": 1390
+    },
+    {
+      "epoch": 0.77,
+      "grad_norm": 0.08351165801286697,
+      "learning_rate": 0.00013531552119442356,
+      "loss": 0.7277,
+      "step": 1391
+    },
+    {
+      "epoch": 0.77,
+      "grad_norm": 0.050557348877191544,
+      "learning_rate": 0.00013523354548476468,
+      "loss": 0.7675,
+      "step": 1392
+    },
+    {
+      "epoch": 0.77,
+      "grad_norm": 0.04991862177848816,
+      "learning_rate": 0.0001351515427333121,
+      "loss": 0.666,
+      "step": 1393
+    },
+    {
+      "epoch": 0.78,
+      "grad_norm": 0.04630931839346886,
+      "learning_rate": 0.0001350695130030031,
+      "loss": 0.6061,
+      "step": 1394
+    },
+    {
+      "epoch": 0.78,
+      "grad_norm": 0.05627691000699997,
+      "learning_rate": 0.00013498745635679557,
+      "loss": 0.7015,
+      "step": 1395
+    },
+    {
+      "epoch": 0.78,
+      "grad_norm": 0.05329517647624016,
+      "learning_rate": 0.00013490537285766808,
+      "loss": 0.6464,
+      "step": 1396
+    },
+    {
+      "epoch": 0.78,
+      "grad_norm": 0.059979844838380814,
+      "learning_rate": 0.00013482326256861988,
+      "loss": 0.6884,
+      "step": 1397
+    },
+    {
+      "epoch": 0.78,
+      "grad_norm": 0.04928671196103096,
+      "learning_rate": 0.0001347411255526707,
+      "loss": 0.6286,
+      "step": 1398
+    },
+    {
+      "epoch": 0.78,
+      "grad_norm": 0.05504937097430229,
+      "learning_rate": 0.00013465896187286083,
+      "loss": 0.7726,
+      "step": 1399
+    },
+    {
+      "epoch": 0.78,
+      "grad_norm": 0.04796033352613449,
+      "learning_rate": 0.00013457677159225097,
+      "loss": 0.5906,
+      "step": 1400
+    },
+    {
+      "epoch": 0.78,
+      "grad_norm": 0.05083645507693291,
+      "learning_rate": 0.0001344945547739223,
+      "loss": 0.6439,
+      "step": 1401
+    },
+    {
+      "epoch": 0.78,
+      "grad_norm": 0.05347789451479912,
+      "learning_rate": 0.0001344123114809763,
+      "loss": 0.6212,
+      "step": 1402
+    },
+    {
+      "epoch": 0.78,
+      "grad_norm": 0.04871118441224098,
+      "learning_rate": 0.00013433004177653486,
+      "loss": 0.6443,
+      "step": 1403
+    },
+    {
+      "epoch": 0.78,
+      "grad_norm": 0.05615173652768135,
+      "learning_rate": 0.00013424774572374005,
+      "loss": 0.7706,
+      "step": 1404
+    },
+    {
+      "epoch": 0.78,
+      "grad_norm": 0.04990394786000252,
+      "learning_rate": 0.00013416542338575424,
+      "loss": 0.5935,
+      "step": 1405
+    },
+    {
+      "epoch": 0.78,
+      "grad_norm": 0.05090782046318054,
+      "learning_rate": 0.0001340830748257599,
+      "loss": 0.6519,
+      "step": 1406
+    },
+    {
+      "epoch": 0.78,
+      "grad_norm": 0.05769623816013336,
+      "learning_rate": 0.00013400070010695966,
+      "loss": 0.6945,
+      "step": 1407
+    },
+    {
+      "epoch": 0.78,
+      "grad_norm": 0.05073278769850731,
+      "learning_rate": 0.00013391829929257625,
+      "loss": 0.6583,
+      "step": 1408
+    },
+    {
+      "epoch": 0.78,
+      "grad_norm": 0.05385829880833626,
+      "learning_rate": 0.0001338358724458524,
+      "loss": 0.6754,
+      "step": 1409
+    },
+    {
+      "epoch": 0.78,
+      "grad_norm": 0.06437458097934723,
+      "learning_rate": 0.0001337534196300508,
+      "loss": 0.7668,
+      "step": 1410
+    },
+    {
+      "epoch": 0.78,
+      "grad_norm": 0.0532778799533844,
+      "learning_rate": 0.0001336709409084542,
+      "loss": 0.6855,
+      "step": 1411
+    },
+    {
+      "epoch": 0.79,
+      "grad_norm": 0.06863218545913696,
+      "learning_rate": 0.00013358843634436496,
+      "loss": 0.7613,
+      "step": 1412
+    },
+    {
+      "epoch": 0.79,
+      "grad_norm": 0.05395227298140526,
+      "learning_rate": 0.00013350590600110556,
+      "loss": 0.6725,
+      "step": 1413
+    },
+    {
+      "epoch": 0.79,
+      "grad_norm": 0.0562172457575798,
+      "learning_rate": 0.00013342334994201815,
+      "loss": 0.7959,
+      "step": 1414
+    },
+    {
+      "epoch": 0.79,
+      "grad_norm": 0.052872899919748306,
+      "learning_rate": 0.00013334076823046456,
+      "loss": 0.666,
+      "step": 1415
+    },
+    {
+      "epoch": 0.79,
+      "grad_norm": 0.05251738429069519,
+      "learning_rate": 0.0001332581609298264,
+      "loss": 0.686,
+      "step": 1416
+    },
+    {
+      "epoch": 0.79,
+      "grad_norm": 0.05102743208408356,
+      "learning_rate": 0.00013317552810350488,
+      "loss": 0.5487,
+      "step": 1417
+    },
+    {
+      "epoch": 0.79,
+      "grad_norm": 0.056528471410274506,
+      "learning_rate": 0.00013309286981492085,
+      "loss": 0.6795,
+      "step": 1418
+    },
+    {
+      "epoch": 0.79,
+      "grad_norm": 0.05416551232337952,
+      "learning_rate": 0.00013301018612751458,
+      "loss": 0.6314,
+      "step": 1419
+    },
+    {
+      "epoch": 0.79,
+      "grad_norm": 0.052541207522153854,
+      "learning_rate": 0.00013292747710474593,
+      "loss": 0.6579,
+      "step": 1420
+    },
+    {
+      "epoch": 0.79,
+      "grad_norm": 0.05116291716694832,
+      "learning_rate": 0.0001328447428100942,
+      "loss": 0.6836,
+      "step": 1421
+    },
+    {
+      "epoch": 0.79,
+      "grad_norm": 0.05055323615670204,
+      "learning_rate": 0.0001327619833070581,
+      "loss": 0.6109,
+      "step": 1422
+    },
+    {
+      "epoch": 0.79,
+      "grad_norm": 0.06023389473557472,
+      "learning_rate": 0.00013267919865915564,
+      "loss": 0.7022,
+      "step": 1423
+    },
+    {
+      "epoch": 0.79,
+      "grad_norm": 0.050478942692279816,
+      "learning_rate": 0.00013259638892992412,
+      "loss": 0.5096,
+      "step": 1424
+    },
+    {
+      "epoch": 0.79,
+      "grad_norm": 0.050790175795555115,
+      "learning_rate": 0.00013251355418292019,
+      "loss": 0.7475,
+      "step": 1425
+    },
+    {
+      "epoch": 0.79,
+      "grad_norm": 0.05030534043908119,
+      "learning_rate": 0.00013243069448171953,
+      "loss": 0.6937,
+      "step": 1426
+    },
+    {
+      "epoch": 0.79,
+      "grad_norm": 0.04698512703180313,
+      "learning_rate": 0.00013234780988991712,
+      "loss": 0.5889,
+      "step": 1427
+    },
+    {
+      "epoch": 0.79,
+      "grad_norm": 0.05166896432638168,
+      "learning_rate": 0.00013226490047112702,
+      "loss": 0.7394,
+      "step": 1428
+    },
+    {
+      "epoch": 0.79,
+      "grad_norm": 0.05660312995314598,
+      "learning_rate": 0.00013218196628898233,
+      "loss": 0.6686,
+      "step": 1429
+    },
+    {
+      "epoch": 0.8,
+      "grad_norm": 0.04474378004670143,
+      "learning_rate": 0.00013209900740713507,
+      "loss": 0.6026,
+      "step": 1430
+    },
+    {
+      "epoch": 0.8,
+      "grad_norm": 0.06732626259326935,
+      "learning_rate": 0.00013201602388925637,
+      "loss": 0.8321,
+      "step": 1431
+    },
+    {
+      "epoch": 0.8,
+      "grad_norm": 0.05723815783858299,
+      "learning_rate": 0.00013193301579903616,
+      "loss": 0.6508,
+      "step": 1432
+    },
+    {
+      "epoch": 0.8,
+      "grad_norm": 0.05704987049102783,
+      "learning_rate": 0.00013184998320018326,
+      "loss": 0.7649,
+      "step": 1433
+    },
+    {
+      "epoch": 0.8,
+      "grad_norm": 0.05435892567038536,
+      "learning_rate": 0.00013176692615642533,
+      "loss": 0.7658,
+      "step": 1434
+    },
+    {
+      "epoch": 0.8,
+      "grad_norm": 0.04823015630245209,
+      "learning_rate": 0.0001316838447315087,
+      "loss": 0.6798,
+      "step": 1435
+    },
+    {
+      "epoch": 0.8,
+      "grad_norm": 0.05567566677927971,
+      "learning_rate": 0.00013160073898919853,
+      "loss": 0.6289,
+      "step": 1436
+    },
+    {
+      "epoch": 0.8,
+      "grad_norm": 0.059023939073085785,
+      "learning_rate": 0.0001315176089932786,
+      "loss": 0.7173,
+      "step": 1437
+    },
+    {
+      "epoch": 0.8,
+      "grad_norm": 0.052320726215839386,
+      "learning_rate": 0.00013143445480755123,
+      "loss": 0.6257,
+      "step": 1438
+    },
+    {
+      "epoch": 0.8,
+      "grad_norm": 0.05305778235197067,
+      "learning_rate": 0.00013135127649583744,
+      "loss": 0.6709,
+      "step": 1439
+    },
+    {
+      "epoch": 0.8,
+      "grad_norm": 0.054861389100551605,
+      "learning_rate": 0.00013126807412197665,
+      "loss": 0.6239,
+      "step": 1440
+    },
+    {
+      "epoch": 0.8,
+      "grad_norm": 0.04839559271931648,
+      "learning_rate": 0.0001311848477498268,
+      "loss": 0.584,
+      "step": 1441
+    },
+    {
+      "epoch": 0.8,
+      "grad_norm": 0.04709853231906891,
+      "learning_rate": 0.00013110159744326427,
+      "loss": 0.5997,
+      "step": 1442
+    },
+    {
+      "epoch": 0.8,
+      "grad_norm": 0.05461667478084564,
+      "learning_rate": 0.00013101832326618376,
+      "loss": 0.6488,
+      "step": 1443
+    },
+    {
+      "epoch": 0.8,
+      "grad_norm": 0.05399639531970024,
+      "learning_rate": 0.0001309350252824983,
+      "loss": 0.6913,
+      "step": 1444
+    },
+    {
+      "epoch": 0.8,
+      "grad_norm": 0.05999092012643814,
+      "learning_rate": 0.00013085170355613926,
+      "loss": 0.6265,
+      "step": 1445
+    },
+    {
+      "epoch": 0.8,
+      "grad_norm": 0.05753492936491966,
+      "learning_rate": 0.0001307683581510561,
+      "loss": 0.7015,
+      "step": 1446
+    },
+    {
+      "epoch": 0.8,
+      "grad_norm": 0.054452694952487946,
+      "learning_rate": 0.00013068498913121657,
+      "loss": 0.6592,
+      "step": 1447
+    },
+    {
+      "epoch": 0.81,
+      "grad_norm": 0.05475104600191116,
+      "learning_rate": 0.00013060159656060654,
+      "loss": 0.6994,
+      "step": 1448
+    },
+    {
+      "epoch": 0.81,
+      "grad_norm": 0.05358335003256798,
+      "learning_rate": 0.00013051818050322986,
+      "loss": 0.6967,
+      "step": 1449
+    },
+    {
+      "epoch": 0.81,
+      "grad_norm": 0.05165765434503555,
+      "learning_rate": 0.0001304347410231085,
+      "loss": 0.681,
+      "step": 1450
+    },
+    {
+      "epoch": 0.81,
+      "grad_norm": 0.05551374331116676,
+      "learning_rate": 0.0001303512781842824,
+      "loss": 0.565,
+      "step": 1451
+    },
+    {
+      "epoch": 0.81,
+      "grad_norm": 0.05121142789721489,
+      "learning_rate": 0.00013026779205080932,
+      "loss": 0.6739,
+      "step": 1452
+    },
+    {
+      "epoch": 0.81,
+      "grad_norm": 0.05909129977226257,
+      "learning_rate": 0.00013018428268676504,
+      "loss": 0.8198,
+      "step": 1453
+    },
+    {
+      "epoch": 0.81,
+      "grad_norm": 0.048282720148563385,
+      "learning_rate": 0.0001301007501562431,
+      "loss": 0.636,
+      "step": 1454
+    },
+    {
+      "epoch": 0.81,
+      "grad_norm": 0.05285617336630821,
+      "learning_rate": 0.00013001719452335485,
+      "loss": 0.623,
+      "step": 1455
+    },
+    {
+      "epoch": 0.81,
+      "grad_norm": 0.05915940925478935,
+      "learning_rate": 0.00012993361585222928,
+      "loss": 0.7059,
+      "step": 1456
+    },
+    {
+      "epoch": 0.81,
+      "grad_norm": 0.0531185120344162,
+      "learning_rate": 0.00012985001420701318,
+      "loss": 0.6381,
+      "step": 1457
+    },
+    {
+      "epoch": 0.81,
+      "grad_norm": 0.05316080525517464,
+      "learning_rate": 0.00012976638965187095,
+      "loss": 0.6969,
+      "step": 1458
+    },
+    {
+      "epoch": 0.81,
+      "grad_norm": 0.052480001002550125,
+      "learning_rate": 0.00012968274225098452,
+      "loss": 0.6696,
+      "step": 1459
+    },
+    {
+      "epoch": 0.81,
+      "grad_norm": 0.05531112849712372,
+      "learning_rate": 0.00012959907206855343,
+      "loss": 0.7064,
+      "step": 1460
+    },
+    {
+      "epoch": 0.81,
+      "grad_norm": 0.055633753538131714,
+      "learning_rate": 0.00012951537916879458,
+      "loss": 0.6559,
+      "step": 1461
+    },
+    {
+      "epoch": 0.81,
+      "grad_norm": 0.052477676421403885,
+      "learning_rate": 0.00012943166361594242,
+      "loss": 0.6901,
+      "step": 1462
+    },
+    {
+      "epoch": 0.81,
+      "grad_norm": 0.05346304923295975,
+      "learning_rate": 0.00012934792547424873,
+      "loss": 0.6428,
+      "step": 1463
+    },
+    {
+      "epoch": 0.81,
+      "grad_norm": 0.05244500935077667,
+      "learning_rate": 0.0001292641648079827,
+      "loss": 0.6615,
+      "step": 1464
+    },
+    {
+      "epoch": 0.81,
+      "grad_norm": 0.05374164506793022,
+      "learning_rate": 0.00012918038168143066,
+      "loss": 0.711,
+      "step": 1465
+    },
+    {
+      "epoch": 0.82,
+      "grad_norm": 0.05638580396771431,
+      "learning_rate": 0.00012909657615889638,
+      "loss": 0.731,
+      "step": 1466
+    },
+    {
+      "epoch": 0.82,
+      "grad_norm": 0.05788502097129822,
+      "learning_rate": 0.00012901274830470064,
+      "loss": 0.7525,
+      "step": 1467
+    },
+    {
+      "epoch": 0.82,
+      "grad_norm": 0.05340276658535004,
+      "learning_rate": 0.0001289288981831815,
+      "loss": 0.6505,
+      "step": 1468
+    },
+    {
+      "epoch": 0.82,
+      "grad_norm": 0.05570358410477638,
+      "learning_rate": 0.00012884502585869395,
+      "loss": 0.6989,
+      "step": 1469
+    },
+    {
+      "epoch": 0.82,
+      "grad_norm": 0.05580015480518341,
+      "learning_rate": 0.00012876113139561018,
+      "loss": 0.7683,
+      "step": 1470
+    },
+    {
+      "epoch": 0.82,
+      "grad_norm": 0.05318528413772583,
+      "learning_rate": 0.0001286772148583193,
+      "loss": 0.6406,
+      "step": 1471
+    },
+    {
+      "epoch": 0.82,
+      "grad_norm": 0.05146130174398422,
+      "learning_rate": 0.0001285932763112273,
+      "loss": 0.65,
+      "step": 1472
+    },
+    {
+      "epoch": 0.82,
+      "grad_norm": 0.05019865557551384,
+      "learning_rate": 0.00012850931581875723,
+      "loss": 0.6883,
+      "step": 1473
+    },
+    {
+      "epoch": 0.82,
+      "grad_norm": 0.06139828637242317,
+      "learning_rate": 0.00012842533344534877,
+      "loss": 0.6625,
+      "step": 1474
+    },
+    {
+      "epoch": 0.82,
+      "grad_norm": 0.055851712822914124,
+      "learning_rate": 0.0001283413292554586,
+      "loss": 0.6335,
+      "step": 1475
+    },
+    {
+      "epoch": 0.82,
+      "grad_norm": 0.05272622033953667,
+      "learning_rate": 0.00012825730331355995,
+      "loss": 0.6381,
+      "step": 1476
+    },
+    {
+      "epoch": 0.82,
+      "grad_norm": 0.053820542991161346,
+      "learning_rate": 0.00012817325568414297,
+      "loss": 0.6784,
+      "step": 1477
+    },
+    {
+      "epoch": 0.82,
+      "grad_norm": 0.05233492702245712,
+      "learning_rate": 0.00012808918643171424,
+      "loss": 0.6492,
+      "step": 1478
+    },
+    {
+      "epoch": 0.82,
+      "grad_norm": 0.05471348017454147,
+      "learning_rate": 0.00012800509562079705,
+      "loss": 0.6332,
+      "step": 1479
+    },
+    {
+      "epoch": 0.82,
+      "grad_norm": 0.051646675914525986,
+      "learning_rate": 0.0001279209833159312,
+      "loss": 0.6563,
+      "step": 1480
+    },
+    {
+      "epoch": 0.82,
+      "grad_norm": 0.05445919558405876,
+      "learning_rate": 0.00012783684958167304,
+      "loss": 0.7054,
+      "step": 1481
+    },
+    {
+      "epoch": 0.82,
+      "grad_norm": 0.053671520203351974,
+      "learning_rate": 0.00012775269448259526,
+      "loss": 0.6493,
+      "step": 1482
+    },
+    {
+      "epoch": 0.82,
+      "grad_norm": 0.05467415601015091,
+      "learning_rate": 0.00012766851808328707,
+      "loss": 0.6223,
+      "step": 1483
+    },
+    {
+      "epoch": 0.83,
+      "grad_norm": 0.05039871484041214,
+      "learning_rate": 0.00012758432044835392,
+      "loss": 0.5863,
+      "step": 1484
+    },
+    {
+      "epoch": 0.83,
+      "grad_norm": 0.048955757170915604,
+      "learning_rate": 0.00012750010164241764,
+      "loss": 0.5957,
+      "step": 1485
+    },
+    {
+      "epoch": 0.83,
+      "grad_norm": 0.06016591191291809,
+      "learning_rate": 0.00012741586173011625,
+      "loss": 0.7174,
+      "step": 1486
+    },
+    {
+      "epoch": 0.83,
+      "grad_norm": 0.05054686218500137,
+      "learning_rate": 0.00012733160077610403,
+      "loss": 0.5833,
+      "step": 1487
+    },
+    {
+      "epoch": 0.83,
+      "grad_norm": 0.06371674686670303,
+      "learning_rate": 0.00012724731884505134,
+      "loss": 0.6756,
+      "step": 1488
+    },
+    {
+      "epoch": 0.83,
+      "grad_norm": 0.04819337651133537,
+      "learning_rate": 0.0001271630160016447,
+      "loss": 0.6621,
+      "step": 1489
+    },
+    {
+      "epoch": 0.83,
+      "grad_norm": 0.053725238889455795,
+      "learning_rate": 0.00012707869231058666,
+      "loss": 0.6667,
+      "step": 1490
+    },
+    {
+      "epoch": 0.83,
+      "grad_norm": 0.04935624450445175,
+      "learning_rate": 0.00012699434783659577,
+      "loss": 0.6873,
+      "step": 1491
+    },
+    {
+      "epoch": 0.83,
+      "grad_norm": 0.05342969298362732,
+      "learning_rate": 0.00012690998264440652,
+      "loss": 0.7205,
+      "step": 1492
+    },
+    {
+      "epoch": 0.83,
+      "grad_norm": 0.04896765202283859,
+      "learning_rate": 0.0001268255967987693,
+      "loss": 0.5972,
+      "step": 1493
+    },
+    {
+      "epoch": 0.83,
+      "grad_norm": 0.05099526047706604,
+      "learning_rate": 0.00012674119036445037,
+      "loss": 0.6677,
+      "step": 1494
+    },
+    {
+      "epoch": 0.83,
+      "grad_norm": 0.0643346831202507,
+      "learning_rate": 0.0001266567634062317,
+      "loss": 0.7039,
+      "step": 1495
+    },
+    {
+      "epoch": 0.83,
+      "grad_norm": 0.06624721735715866,
+      "learning_rate": 0.00012657231598891126,
+      "loss": 0.7797,
+      "step": 1496
+    },
+    {
+      "epoch": 0.83,
+      "grad_norm": 0.05854468047618866,
+      "learning_rate": 0.00012648784817730242,
+      "loss": 0.7198,
+      "step": 1497
+    },
+    {
+      "epoch": 0.83,
+      "grad_norm": 0.05921991169452667,
+      "learning_rate": 0.00012640336003623444,
+      "loss": 0.7499,
+      "step": 1498
+    },
+    {
+      "epoch": 0.83,
+      "grad_norm": 0.055420514196157455,
+      "learning_rate": 0.000126318851630552,
+      "loss": 0.5993,
+      "step": 1499
+    },
+    {
+      "epoch": 0.83,
+      "grad_norm": 0.06067932769656181,
+      "learning_rate": 0.00012623432302511544,
+      "loss": 0.8006,
+      "step": 1500
+    },
+    {
+      "epoch": 0.83,
+      "grad_norm": 0.05395696684718132,
+      "learning_rate": 0.0001261497742848006,
+      "loss": 0.654,
+      "step": 1501
+    },
+    {
+      "epoch": 0.84,
+      "grad_norm": 0.05506988614797592,
+      "learning_rate": 0.0001260652054744987,
+      "loss": 0.6884,
+      "step": 1502
+    },
+    {
+      "epoch": 0.84,
+      "grad_norm": 0.0567973367869854,
+      "learning_rate": 0.0001259806166591165,
+      "loss": 0.685,
+      "step": 1503
+    },
+    {
+      "epoch": 0.84,
+      "grad_norm": 0.05710934102535248,
+      "learning_rate": 0.00012589600790357592,
+      "loss": 0.6786,
+      "step": 1504
+    },
+    {
+      "epoch": 0.84,
+      "grad_norm": 0.0530974306166172,
+      "learning_rate": 0.00012581137927281439,
+      "loss": 0.6175,
+      "step": 1505
+    },
+    {
+      "epoch": 0.84,
+      "grad_norm": 0.05763082951307297,
+      "learning_rate": 0.0001257267308317845,
+      "loss": 0.5977,
+      "step": 1506
+    },
+    {
+      "epoch": 0.84,
+      "grad_norm": 0.056460849940776825,
+      "learning_rate": 0.00012564206264545396,
+      "loss": 0.706,
+      "step": 1507
+    },
+    {
+      "epoch": 0.84,
+      "grad_norm": 0.05578916147351265,
+      "learning_rate": 0.00012555737477880577,
+      "loss": 0.736,
+      "step": 1508
+    },
+    {
+      "epoch": 0.84,
+      "grad_norm": 0.05610078200697899,
+      "learning_rate": 0.00012547266729683797,
+      "loss": 0.6614,
+      "step": 1509
+    },
+    {
+      "epoch": 0.84,
+      "grad_norm": 0.05859774723649025,
+      "learning_rate": 0.00012538794026456366,
+      "loss": 0.7192,
+      "step": 1510
+    },
+    {
+      "epoch": 0.84,
+      "grad_norm": 0.05389096215367317,
+      "learning_rate": 0.00012530319374701098,
+      "loss": 0.6748,
+      "step": 1511
+    },
+    {
+      "epoch": 0.84,
+      "grad_norm": 0.05462174117565155,
+      "learning_rate": 0.000125218427809223,
+      "loss": 0.6765,
+      "step": 1512
+    },
+    {
+      "epoch": 0.84,
+      "grad_norm": 0.047790851444005966,
+      "learning_rate": 0.00012513364251625766,
+      "loss": 0.6126,
+      "step": 1513
+    },
+    {
+      "epoch": 0.84,
+      "grad_norm": 0.05302601680159569,
+      "learning_rate": 0.0001250488379331878,
+      "loss": 0.6352,
+      "step": 1514
+    },
+    {
+      "epoch": 0.84,
+      "grad_norm": 0.05396215245127678,
+      "learning_rate": 0.00012496401412510102,
+      "loss": 0.6203,
+      "step": 1515
+    },
+    {
+      "epoch": 0.84,
+      "grad_norm": 0.051495350897312164,
+      "learning_rate": 0.00012487917115709975,
+      "loss": 0.7411,
+      "step": 1516
+    },
+    {
+      "epoch": 0.84,
+      "grad_norm": 0.05668788030743599,
+      "learning_rate": 0.00012479430909430108,
+      "loss": 0.7574,
+      "step": 1517
+    },
+    {
+      "epoch": 0.84,
+      "grad_norm": 0.056466199457645416,
+      "learning_rate": 0.00012470942800183675,
+      "loss": 0.6395,
+      "step": 1518
+    },
+    {
+      "epoch": 0.84,
+      "grad_norm": 0.061602648347616196,
+      "learning_rate": 0.0001246245279448531,
+      "loss": 0.7101,
+      "step": 1519
+    },
+    {
+      "epoch": 0.85,
+      "grad_norm": 0.05588141828775406,
+      "learning_rate": 0.00012453960898851108,
+      "loss": 0.7231,
+      "step": 1520
+    },
+    {
+      "epoch": 0.85,
+      "grad_norm": 0.05210031941533089,
+      "learning_rate": 0.00012445467119798605,
+      "loss": 0.7094,
+      "step": 1521
+    },
+    {
+      "epoch": 0.85,
+      "grad_norm": 0.05762970820069313,
+      "learning_rate": 0.00012436971463846788,
+      "loss": 0.6623,
+      "step": 1522
+    },
+    {
+      "epoch": 0.85,
+      "grad_norm": 0.05559748038649559,
+      "learning_rate": 0.0001242847393751609,
+      "loss": 0.6677,
+      "step": 1523
+    },
+    {
+      "epoch": 0.85,
+      "grad_norm": 0.05742710828781128,
+      "learning_rate": 0.00012419974547328366,
+      "loss": 0.7276,
+      "step": 1524
+    },
+    {
+      "epoch": 0.85,
+      "grad_norm": 0.053973183035850525,
+      "learning_rate": 0.00012411473299806918,
+      "loss": 0.7039,
+      "step": 1525
+    },
+    {
+      "epoch": 0.85,
+      "grad_norm": 0.056114938110113144,
+      "learning_rate": 0.00012402970201476457,
+      "loss": 0.6718,
+      "step": 1526
+    },
+    {
+      "epoch": 0.85,
+      "grad_norm": 0.06136467307806015,
+      "learning_rate": 0.0001239446525886313,
+      "loss": 0.7666,
+      "step": 1527
+    },
+    {
+      "epoch": 0.85,
+      "grad_norm": 0.057720523327589035,
+      "learning_rate": 0.00012385958478494487,
+      "loss": 0.6303,
+      "step": 1528
+    },
+    {
+      "epoch": 0.85,
+      "grad_norm": 0.054988570511341095,
+      "learning_rate": 0.00012377449866899493,
+      "loss": 0.6074,
+      "step": 1529
+    },
+    {
+      "epoch": 0.85,
+      "grad_norm": 0.05281984433531761,
+      "learning_rate": 0.00012368939430608522,
+      "loss": 0.6453,
+      "step": 1530
+    },
+    {
+      "epoch": 0.85,
+      "grad_norm": 0.06022608280181885,
+      "learning_rate": 0.00012360427176153342,
+      "loss": 0.7284,
+      "step": 1531
+    },
+    {
+      "epoch": 0.85,
+      "grad_norm": 0.05347907170653343,
+      "learning_rate": 0.00012351913110067122,
+      "loss": 0.5784,
+      "step": 1532
+    },
+    {
+      "epoch": 0.85,
+      "grad_norm": 0.0521782748401165,
+      "learning_rate": 0.0001234339723888442,
+      "loss": 0.6836,
+      "step": 1533
+    },
+    {
+      "epoch": 0.85,
+      "grad_norm": 0.05518525093793869,
+      "learning_rate": 0.00012334879569141172,
+      "loss": 0.6714,
+      "step": 1534
+    },
+    {
+      "epoch": 0.85,
+      "grad_norm": 0.0590486079454422,
+      "learning_rate": 0.00012326360107374712,
+      "loss": 0.7145,
+      "step": 1535
+    },
+    {
+      "epoch": 0.85,
+      "grad_norm": 0.05373707413673401,
+      "learning_rate": 0.0001231783886012373,
+      "loss": 0.6897,
+      "step": 1536
+    },
+    {
+      "epoch": 0.85,
+      "grad_norm": 0.05092402547597885,
+      "learning_rate": 0.00012309315833928302,
+      "loss": 0.6245,
+      "step": 1537
+    },
+    {
+      "epoch": 0.86,
+      "grad_norm": 0.05266068875789642,
+      "learning_rate": 0.00012300791035329853,
+      "loss": 0.5606,
+      "step": 1538
+    },
+    {
+      "epoch": 0.86,
+      "grad_norm": 0.05256986245512962,
+      "learning_rate": 0.00012292264470871182,
+      "loss": 0.6055,
+      "step": 1539
+    },
+    {
+      "epoch": 0.86,
+      "grad_norm": 0.05765622481703758,
+      "learning_rate": 0.0001228373614709644,
+      "loss": 0.7045,
+      "step": 1540
+    },
+    {
+      "epoch": 0.86,
+      "grad_norm": 0.05402863398194313,
+      "learning_rate": 0.00012275206070551126,
+      "loss": 0.7477,
+      "step": 1541
+    },
+    {
+      "epoch": 0.86,
+      "grad_norm": 0.049794167280197144,
+      "learning_rate": 0.00012266674247782085,
+      "loss": 0.6824,
+      "step": 1542
+    },
+    {
+      "epoch": 0.86,
+      "grad_norm": 0.052845150232315063,
+      "learning_rate": 0.000122581406853375,
+      "loss": 0.6614,
+      "step": 1543
+    },
+    {
+      "epoch": 0.86,
+      "grad_norm": 0.05592498928308487,
+      "learning_rate": 0.00012249605389766895,
+      "loss": 0.6355,
+      "step": 1544
+    },
+    {
+      "epoch": 0.86,
+      "grad_norm": 0.05580741912126541,
+      "learning_rate": 0.0001224106836762112,
+      "loss": 0.557,
+      "step": 1545
+    },
+    {
+      "epoch": 0.86,
+      "grad_norm": 0.05675602704286575,
+      "learning_rate": 0.00012232529625452352,
+      "loss": 0.6461,
+      "step": 1546
+    },
+    {
+      "epoch": 0.86,
+      "grad_norm": 0.05478064343333244,
+      "learning_rate": 0.00012223989169814087,
+      "loss": 0.6754,
+      "step": 1547
+    },
+    {
+      "epoch": 0.86,
+      "grad_norm": 0.05122774466872215,
+      "learning_rate": 0.00012215447007261134,
+      "loss": 0.6126,
+      "step": 1548
+    },
+    {
+      "epoch": 0.86,
+      "grad_norm": 0.05637194588780403,
+      "learning_rate": 0.00012206903144349615,
+      "loss": 0.599,
+      "step": 1549
+    },
+    {
+      "epoch": 0.86,
+      "grad_norm": 0.06174594536423683,
+      "learning_rate": 0.00012198357587636957,
+      "loss": 0.7692,
+      "step": 1550
+    },
+    {
+      "epoch": 0.86,
+      "grad_norm": 0.0552188940346241,
+      "learning_rate": 0.00012189810343681889,
+      "loss": 0.5888,
+      "step": 1551
+    },
+    {
+      "epoch": 0.86,
+      "grad_norm": 0.06255478411912918,
+      "learning_rate": 0.00012181261419044428,
+      "loss": 0.6878,
+      "step": 1552
+    },
+    {
+      "epoch": 0.86,
+      "grad_norm": 0.049522750079631805,
+      "learning_rate": 0.00012172710820285885,
+      "loss": 0.645,
+      "step": 1553
+    },
+    {
+      "epoch": 0.86,
+      "grad_norm": 0.05213455855846405,
+      "learning_rate": 0.00012164158553968856,
+      "loss": 0.5865,
+      "step": 1554
+    },
+    {
+      "epoch": 0.86,
+      "grad_norm": 0.05377736687660217,
+      "learning_rate": 0.00012155604626657222,
+      "loss": 0.6835,
+      "step": 1555
+    },
+    {
+      "epoch": 0.87,
+      "grad_norm": 0.06013650447130203,
+      "learning_rate": 0.0001214704904491613,
+      "loss": 0.6399,
+      "step": 1556
+    },
+    {
+      "epoch": 0.87,
+      "grad_norm": 0.057262714952230453,
+      "learning_rate": 0.00012138491815312001,
+      "loss": 0.68,
+      "step": 1557
+    },
+    {
+      "epoch": 0.87,
+      "grad_norm": 0.05029933899641037,
+      "learning_rate": 0.00012129932944412518,
+      "loss": 0.6265,
+      "step": 1558
+    },
+    {
+      "epoch": 0.87,
+      "grad_norm": 0.05597386881709099,
+      "learning_rate": 0.00012121372438786631,
+      "loss": 0.7077,
+      "step": 1559
+    },
+    {
+      "epoch": 0.87,
+      "grad_norm": 0.05315025895833969,
+      "learning_rate": 0.00012112810305004535,
+      "loss": 0.6948,
+      "step": 1560
+    },
+    {
+      "epoch": 0.87,
+      "grad_norm": 0.056836340576410294,
+      "learning_rate": 0.00012104246549637683,
+      "loss": 0.6079,
+      "step": 1561
+    },
+    {
+      "epoch": 0.87,
+      "grad_norm": 0.05192887783050537,
+      "learning_rate": 0.00012095681179258765,
+      "loss": 0.6264,
+      "step": 1562
+    },
+    {
+      "epoch": 0.87,
+      "grad_norm": 0.05339023098349571,
+      "learning_rate": 0.00012087114200441714,
+      "loss": 0.6584,
+      "step": 1563
+    },
+    {
+      "epoch": 0.87,
+      "grad_norm": 0.054986581206321716,
+      "learning_rate": 0.00012078545619761703,
+      "loss": 0.5941,
+      "step": 1564
+    },
+    {
+      "epoch": 0.87,
+      "grad_norm": 0.05665560066699982,
+      "learning_rate": 0.00012069975443795126,
+      "loss": 0.6785,
+      "step": 1565
+    },
+    {
+      "epoch": 0.87,
+      "grad_norm": 0.055501628667116165,
+      "learning_rate": 0.00012061403679119603,
+      "loss": 0.7018,
+      "step": 1566
+    },
+    {
+      "epoch": 0.87,
+      "grad_norm": 0.058468155562877655,
+      "learning_rate": 0.00012052830332313978,
+      "loss": 0.76,
+      "step": 1567
+    },
+    {
+      "epoch": 0.87,
+      "grad_norm": 0.055983930826187134,
+      "learning_rate": 0.00012044255409958305,
+      "loss": 0.7121,
+      "step": 1568
+    },
+    {
+      "epoch": 0.87,
+      "grad_norm": 0.056311774998903275,
+      "learning_rate": 0.00012035678918633848,
+      "loss": 0.6893,
+      "step": 1569
+    },
+    {
+      "epoch": 0.87,
+      "grad_norm": 0.05754856392741203,
+      "learning_rate": 0.00012027100864923076,
+      "loss": 0.7448,
+      "step": 1570
+    },
+    {
+      "epoch": 0.87,
+      "grad_norm": 0.05376417189836502,
+      "learning_rate": 0.00012018521255409656,
+      "loss": 0.667,
+      "step": 1571
+    },
+    {
+      "epoch": 0.87,
+      "grad_norm": 0.0655338391661644,
+      "learning_rate": 0.00012009940096678452,
+      "loss": 0.7666,
+      "step": 1572
+    },
+    {
+      "epoch": 0.87,
+      "grad_norm": 0.055434565991163254,
+      "learning_rate": 0.00012001357395315511,
+      "loss": 0.5904,
+      "step": 1573
+    },
+    {
+      "epoch": 0.88,
+      "grad_norm": 0.053381387144327164,
+      "learning_rate": 0.00011992773157908073,
+      "loss": 0.6542,
+      "step": 1574
+    },
+    {
+      "epoch": 0.88,
+      "grad_norm": 0.05259239673614502,
+      "learning_rate": 0.00011984187391044548,
+      "loss": 0.6848,
+      "step": 1575
+    },
+    {
+      "epoch": 0.88,
+      "grad_norm": 0.05607330799102783,
+      "learning_rate": 0.00011975600101314525,
+      "loss": 0.7138,
+      "step": 1576
+    },
+    {
+      "epoch": 0.88,
+      "grad_norm": 0.051935825496912,
+      "learning_rate": 0.00011967011295308761,
+      "loss": 0.6423,
+      "step": 1577
+    },
+    {
+      "epoch": 0.88,
+      "grad_norm": 0.058591440320014954,
+      "learning_rate": 0.00011958420979619176,
+      "loss": 0.6529,
+      "step": 1578
+    },
+    {
+      "epoch": 0.88,
+      "grad_norm": 0.06154406815767288,
+      "learning_rate": 0.00011949829160838844,
+      "loss": 0.7658,
+      "step": 1579
+    },
+    {
+      "epoch": 0.88,
+      "grad_norm": 0.05416911095380783,
+      "learning_rate": 0.00011941235845562006,
+      "loss": 0.6634,
+      "step": 1580
+    },
+    {
+      "epoch": 0.88,
+      "grad_norm": 0.056073009967803955,
+      "learning_rate": 0.00011932641040384038,
+      "loss": 0.575,
+      "step": 1581
+    },
+    {
+      "epoch": 0.88,
+      "grad_norm": 0.06217704340815544,
+      "learning_rate": 0.00011924044751901466,
+      "loss": 0.7539,
+      "step": 1582
+    },
+    {
+      "epoch": 0.88,
+      "grad_norm": 0.05384403467178345,
+      "learning_rate": 0.00011915446986711953,
+      "loss": 0.566,
+      "step": 1583
+    },
+    {
+      "epoch": 0.88,
+      "grad_norm": 0.05363078787922859,
+      "learning_rate": 0.00011906847751414291,
+      "loss": 0.6722,
+      "step": 1584
+    },
+    {
+      "epoch": 0.88,
+      "grad_norm": 0.05821472778916359,
+      "learning_rate": 0.00011898247052608414,
+      "loss": 0.685,
+      "step": 1585
+    },
+    {
+      "epoch": 0.88,
+      "grad_norm": 0.06892435252666473,
+      "learning_rate": 0.00011889644896895362,
+      "loss": 0.6482,
+      "step": 1586
+    },
+    {
+      "epoch": 0.88,
+      "grad_norm": 0.06398369371891022,
+      "learning_rate": 0.00011881041290877303,
+      "loss": 0.675,
+      "step": 1587
+    },
+    {
+      "epoch": 0.88,
+      "grad_norm": 0.05386562645435333,
+      "learning_rate": 0.00011872436241157518,
+      "loss": 0.6116,
+      "step": 1588
+    },
+    {
+      "epoch": 0.88,
+      "grad_norm": 0.05238373950123787,
+      "learning_rate": 0.00011863829754340395,
+      "loss": 0.628,
+      "step": 1589
+    },
+    {
+      "epoch": 0.88,
+      "grad_norm": 0.05979369580745697,
+      "learning_rate": 0.00011855221837031418,
+      "loss": 0.723,
+      "step": 1590
+    },
+    {
+      "epoch": 0.88,
+      "grad_norm": 0.06569157540798187,
+      "learning_rate": 0.00011846612495837182,
+      "loss": 0.7057,
+      "step": 1591
+    },
+    {
+      "epoch": 0.89,
+      "grad_norm": 0.057322900742292404,
+      "learning_rate": 0.00011838001737365365,
+      "loss": 0.6529,
+      "step": 1592
+    },
+    {
+      "epoch": 0.89,
+      "grad_norm": 0.06029359623789787,
+      "learning_rate": 0.00011829389568224734,
+      "loss": 0.7157,
+      "step": 1593
+    },
+    {
+      "epoch": 0.89,
+      "grad_norm": 0.05852045491337776,
+      "learning_rate": 0.00011820775995025147,
+      "loss": 0.6289,
+      "step": 1594
+    },
+    {
+      "epoch": 0.89,
+      "grad_norm": 0.05572306364774704,
+      "learning_rate": 0.00011812161024377526,
+      "loss": 0.6662,
+      "step": 1595
+    },
+    {
+      "epoch": 0.89,
+      "grad_norm": 0.05891454592347145,
+      "learning_rate": 0.00011803544662893875,
+      "loss": 0.662,
+      "step": 1596
+    },
+    {
+      "epoch": 0.89,
+      "grad_norm": 0.05507583171129227,
+      "learning_rate": 0.00011794926917187264,
+      "loss": 0.572,
+      "step": 1597
+    },
+    {
+      "epoch": 0.89,
+      "grad_norm": 0.05523572862148285,
+      "learning_rate": 0.00011786307793871824,
+      "loss": 0.6808,
+      "step": 1598
+    },
+    {
+      "epoch": 0.89,
+      "grad_norm": 0.06197218969464302,
+      "learning_rate": 0.00011777687299562743,
+      "loss": 0.6696,
+      "step": 1599
+    },
+    {
+      "epoch": 0.89,
+      "grad_norm": 0.05204472318291664,
+      "learning_rate": 0.00011769065440876263,
+      "loss": 0.6031,
+      "step": 1600
+    },
+    {
+      "epoch": 0.89,
+      "grad_norm": 0.05904833972454071,
+      "learning_rate": 0.0001176044222442967,
+      "loss": 0.6351,
+      "step": 1601
+    },
+    {
+      "epoch": 0.89,
+      "grad_norm": 0.06034965440630913,
+      "learning_rate": 0.00011751817656841298,
+      "loss": 0.7582,
+      "step": 1602
+    },
+    {
+      "epoch": 0.89,
+      "grad_norm": 0.05305292084813118,
+      "learning_rate": 0.0001174319174473051,
+      "loss": 0.6891,
+      "step": 1603
+    },
+    {
+      "epoch": 0.89,
+      "grad_norm": 0.052683088928461075,
+      "learning_rate": 0.00011734564494717711,
+      "loss": 0.6354,
+      "step": 1604
+    },
+    {
+      "epoch": 0.89,
+      "grad_norm": 0.059987373650074005,
+      "learning_rate": 0.0001172593591342432,
+      "loss": 0.6069,
+      "step": 1605
+    },
+    {
+      "epoch": 0.89,
+      "grad_norm": 0.05370206758379936,
+      "learning_rate": 0.0001171730600747279,
+      "loss": 0.6376,
+      "step": 1606
+    },
+    {
+      "epoch": 0.89,
+      "grad_norm": 0.05009806901216507,
+      "learning_rate": 0.00011708674783486583,
+      "loss": 0.537,
+      "step": 1607
+    },
+    {
+      "epoch": 0.89,
+      "grad_norm": 0.05144254118204117,
+      "learning_rate": 0.00011700042248090176,
+      "loss": 0.6138,
+      "step": 1608
+    },
+    {
+      "epoch": 0.89,
+      "grad_norm": 0.05199851468205452,
+      "learning_rate": 0.0001169140840790905,
+      "loss": 0.6005,
+      "step": 1609
+    },
+    {
+      "epoch": 0.9,
+      "grad_norm": 0.06040224805474281,
+      "learning_rate": 0.00011682773269569693,
+      "loss": 0.6571,
+      "step": 1610
+    },
+    {
+      "epoch": 0.9,
+      "grad_norm": 0.05754357576370239,
+      "learning_rate": 0.00011674136839699581,
+      "loss": 0.7045,
+      "step": 1611
+    },
+    {
+      "epoch": 0.9,
+      "grad_norm": 0.05448243021965027,
+      "learning_rate": 0.00011665499124927184,
+      "loss": 0.642,
+      "step": 1612
+    },
+    {
+      "epoch": 0.9,
+      "grad_norm": 0.0632404014468193,
+      "learning_rate": 0.00011656860131881966,
+      "loss": 0.7868,
+      "step": 1613
+    },
+    {
+      "epoch": 0.9,
+      "grad_norm": 0.054406315088272095,
+      "learning_rate": 0.00011648219867194362,
+      "loss": 0.6224,
+      "step": 1614
+    },
+    {
+      "epoch": 0.9,
+      "grad_norm": 0.059626247733831406,
+      "learning_rate": 0.00011639578337495787,
+      "loss": 0.6865,
+      "step": 1615
+    },
+    {
+      "epoch": 0.9,
+      "grad_norm": 0.05667116865515709,
+      "learning_rate": 0.00011630935549418627,
+      "loss": 0.5988,
+      "step": 1616
+    },
+    {
+      "epoch": 0.9,
+      "grad_norm": 0.0557304285466671,
+      "learning_rate": 0.00011622291509596234,
+      "loss": 0.6422,
+      "step": 1617
+    },
+    {
+      "epoch": 0.9,
+      "grad_norm": 0.0590471550822258,
+      "learning_rate": 0.00011613646224662921,
+      "loss": 0.777,
+      "step": 1618
+    },
+    {
+      "epoch": 0.9,
+      "grad_norm": 0.06502439081668854,
+      "learning_rate": 0.00011604999701253953,
+      "loss": 0.7052,
+      "step": 1619
+    },
+    {
+      "epoch": 0.9,
+      "grad_norm": 0.05545181408524513,
+      "learning_rate": 0.00011596351946005552,
+      "loss": 0.6561,
+      "step": 1620
+    },
+    {
+      "epoch": 0.9,
+      "grad_norm": 0.05329308658838272,
+      "learning_rate": 0.00011587702965554878,
+      "loss": 0.5068,
+      "step": 1621
+    },
+    {
+      "epoch": 0.9,
+      "grad_norm": 0.05559957027435303,
+      "learning_rate": 0.00011579052766540039,
+      "loss": 0.6479,
+      "step": 1622
+    },
+    {
+      "epoch": 0.9,
+      "grad_norm": 0.061965446919202805,
+      "learning_rate": 0.00011570401355600071,
+      "loss": 0.7318,
+      "step": 1623
+    },
+    {
+      "epoch": 0.9,
+      "grad_norm": 0.061798419803380966,
+      "learning_rate": 0.00011561748739374945,
+      "loss": 0.7131,
+      "step": 1624
+    },
+    {
+      "epoch": 0.9,
+      "grad_norm": 0.05673898011445999,
+      "learning_rate": 0.00011553094924505557,
+      "loss": 0.7431,
+      "step": 1625
+    },
+    {
+      "epoch": 0.9,
+      "grad_norm": 0.055540382862091064,
+      "learning_rate": 0.00011544439917633718,
+      "loss": 0.6852,
+      "step": 1626
+    },
+    {
+      "epoch": 0.9,
+      "grad_norm": 0.05674376338720322,
+      "learning_rate": 0.00011535783725402163,
+      "loss": 0.7056,
+      "step": 1627
+    },
+    {
+      "epoch": 0.91,
+      "grad_norm": 0.057587672024965286,
+      "learning_rate": 0.00011527126354454525,
+      "loss": 0.6101,
+      "step": 1628
+    },
+    {
+      "epoch": 0.91,
+      "grad_norm": 0.06215091049671173,
+      "learning_rate": 0.00011518467811435352,
+      "loss": 0.742,
+      "step": 1629
+    },
+    {
+      "epoch": 0.91,
+      "grad_norm": 0.054263584315776825,
+      "learning_rate": 0.00011509808102990085,
+      "loss": 0.6079,
+      "step": 1630
+    },
+    {
+      "epoch": 0.91,
+      "grad_norm": 0.05362584441900253,
+      "learning_rate": 0.00011501147235765063,
+      "loss": 0.5511,
+      "step": 1631
+    },
+    {
+      "epoch": 0.91,
+      "grad_norm": 0.05936874449253082,
+      "learning_rate": 0.00011492485216407513,
+      "loss": 0.6716,
+      "step": 1632
+    },
+    {
+      "epoch": 0.91,
+      "grad_norm": 0.05744662880897522,
+      "learning_rate": 0.00011483822051565549,
+      "loss": 0.6429,
+      "step": 1633
+    },
+    {
+      "epoch": 0.91,
+      "grad_norm": 0.06137154996395111,
+      "learning_rate": 0.0001147515774788816,
+      "loss": 0.6882,
+      "step": 1634
+    },
+    {
+      "epoch": 0.91,
+      "grad_norm": 0.06001037731766701,
+      "learning_rate": 0.0001146649231202521,
+      "loss": 0.655,
+      "step": 1635
+    },
+    {
+      "epoch": 0.91,
+      "grad_norm": 0.054453328251838684,
+      "learning_rate": 0.0001145782575062743,
+      "loss": 0.6446,
+      "step": 1636
+    },
+    {
+      "epoch": 0.91,
+      "grad_norm": 0.056741863489151,
+      "learning_rate": 0.00011449158070346424,
+      "loss": 0.6569,
+      "step": 1637
+    },
+    {
+      "epoch": 0.91,
+      "grad_norm": 0.05473573878407478,
+      "learning_rate": 0.00011440489277834645,
+      "loss": 0.6475,
+      "step": 1638
+    },
+    {
+      "epoch": 0.91,
+      "grad_norm": 0.060271989554166794,
+      "learning_rate": 0.00011431819379745401,
+      "loss": 0.7187,
+      "step": 1639
+    },
+    {
+      "epoch": 0.91,
+      "grad_norm": 0.05573005974292755,
+      "learning_rate": 0.00011423148382732853,
+      "loss": 0.639,
+      "step": 1640
+    },
+    {
+      "epoch": 0.91,
+      "grad_norm": 0.05600711703300476,
+      "learning_rate": 0.00011414476293452001,
+      "loss": 0.6168,
+      "step": 1641
+    },
+    {
+      "epoch": 0.91,
+      "grad_norm": 0.053048767149448395,
+      "learning_rate": 0.0001140580311855869,
+      "loss": 0.5581,
+      "step": 1642
+    },
+    {
+      "epoch": 0.91,
+      "grad_norm": 0.05543733760714531,
+      "learning_rate": 0.00011397128864709586,
+      "loss": 0.6627,
+      "step": 1643
+    },
+    {
+      "epoch": 0.91,
+      "grad_norm": 0.05584166571497917,
+      "learning_rate": 0.00011388453538562195,
+      "loss": 0.6437,
+      "step": 1644
+    },
+    {
+      "epoch": 0.91,
+      "grad_norm": 0.06031232327222824,
+      "learning_rate": 0.00011379777146774844,
+      "loss": 0.5947,
+      "step": 1645
+    },
+    {
+      "epoch": 0.92,
+      "grad_norm": 0.055425215512514114,
+      "learning_rate": 0.0001137109969600667,
+      "loss": 0.6921,
+      "step": 1646
+    },
+    {
+      "epoch": 0.92,
+      "grad_norm": 0.05608774349093437,
+      "learning_rate": 0.00011362421192917631,
+      "loss": 0.7744,
+      "step": 1647
+    },
+    {
+      "epoch": 0.92,
+      "grad_norm": 0.06419660151004791,
+      "learning_rate": 0.00011353741644168487,
+      "loss": 0.7103,
+      "step": 1648
+    },
+    {
+      "epoch": 0.92,
+      "grad_norm": 0.060391318053007126,
+      "learning_rate": 0.0001134506105642081,
+      "loss": 0.7168,
+      "step": 1649
+    },
+    {
+      "epoch": 0.92,
+      "grad_norm": 0.05262453481554985,
+      "learning_rate": 0.00011336379436336955,
+      "loss": 0.5733,
+      "step": 1650
+    },
+    {
+      "epoch": 0.92,
+      "grad_norm": 0.0592212975025177,
+      "learning_rate": 0.00011327696790580083,
+      "loss": 0.6959,
+      "step": 1651
+    },
+    {
+      "epoch": 0.92,
+      "grad_norm": 0.061455368995666504,
+      "learning_rate": 0.00011319013125814131,
+      "loss": 0.661,
+      "step": 1652
+    },
+    {
+      "epoch": 0.92,
+      "grad_norm": 0.05837690457701683,
+      "learning_rate": 0.00011310328448703829,
+      "loss": 0.6521,
+      "step": 1653
+    },
+    {
+      "epoch": 0.92,
+      "grad_norm": 0.055113084614276886,
+      "learning_rate": 0.00011301642765914673,
+      "loss": 0.7173,
+      "step": 1654
+    },
+    {
+      "epoch": 0.92,
+      "grad_norm": 0.06535536795854568,
+      "learning_rate": 0.00011292956084112943,
+      "loss": 0.6889,
+      "step": 1655
+    },
+    {
+      "epoch": 0.92,
+      "grad_norm": 0.056155428290367126,
+      "learning_rate": 0.00011284268409965673,
+      "loss": 0.6287,
+      "step": 1656
+    },
+    {
+      "epoch": 0.92,
+      "grad_norm": 0.05229957774281502,
+      "learning_rate": 0.00011275579750140666,
+      "loss": 0.617,
+      "step": 1657
+    },
+    {
+      "epoch": 0.92,
+      "grad_norm": 0.05848075821995735,
+      "learning_rate": 0.00011266890111306484,
+      "loss": 0.6121,
+      "step": 1658
+    },
+    {
+      "epoch": 0.92,
+      "grad_norm": 0.06387878954410553,
+      "learning_rate": 0.00011258199500132429,
+      "loss": 0.6307,
+      "step": 1659
+    },
+    {
+      "epoch": 0.92,
+      "grad_norm": 0.05076488479971886,
+      "learning_rate": 0.00011249507923288562,
+      "loss": 0.6157,
+      "step": 1660
+    },
+    {
+      "epoch": 0.92,
+      "grad_norm": 0.053368885070085526,
+      "learning_rate": 0.0001124081538744568,
+      "loss": 0.675,
+      "step": 1661
+    },
+    {
+      "epoch": 0.92,
+      "grad_norm": 0.06017875671386719,
+      "learning_rate": 0.00011232121899275314,
+      "loss": 0.6557,
+      "step": 1662
+    },
+    {
+      "epoch": 0.92,
+      "grad_norm": 0.06376302242279053,
+      "learning_rate": 0.00011223427465449729,
+      "loss": 0.6831,
+      "step": 1663
+    },
+    {
+      "epoch": 0.93,
+      "grad_norm": 0.06142215430736542,
+      "learning_rate": 0.00011214732092641916,
+      "loss": 0.6506,
+      "step": 1664
+    },
+    {
+      "epoch": 0.93,
+      "grad_norm": 0.0578768253326416,
+      "learning_rate": 0.00011206035787525585,
+      "loss": 0.7647,
+      "step": 1665
+    },
+    {
+      "epoch": 0.93,
+      "grad_norm": 0.05308634787797928,
+      "learning_rate": 0.00011197338556775156,
+      "loss": 0.6295,
+      "step": 1666
+    },
+    {
+      "epoch": 0.93,
+      "grad_norm": 0.050338249653577805,
+      "learning_rate": 0.00011188640407065776,
+      "loss": 0.5454,
+      "step": 1667
+    },
+    {
+      "epoch": 0.93,
+      "grad_norm": 0.05560674890875816,
+      "learning_rate": 0.00011179941345073278,
+      "loss": 0.7058,
+      "step": 1668
+    },
+    {
+      "epoch": 0.93,
+      "grad_norm": 0.054467298090457916,
+      "learning_rate": 0.00011171241377474207,
+      "loss": 0.5801,
+      "step": 1669
+    },
+    {
+      "epoch": 0.93,
+      "grad_norm": 0.05152300372719765,
+      "learning_rate": 0.00011162540510945799,
+      "loss": 0.5644,
+      "step": 1670
+    },
+    {
+      "epoch": 0.93,
+      "grad_norm": 0.05688504874706268,
+      "learning_rate": 0.0001115383875216598,
+      "loss": 0.6589,
+      "step": 1671
+    },
+    {
+      "epoch": 0.93,
+      "grad_norm": 0.05523272603750229,
+      "learning_rate": 0.00011145136107813363,
+      "loss": 0.6098,
+      "step": 1672
+    },
+    {
+      "epoch": 0.93,
+      "grad_norm": 0.058240048587322235,
+      "learning_rate": 0.0001113643258456724,
+      "loss": 0.7046,
+      "step": 1673
+    },
+    {
+      "epoch": 0.93,
+      "grad_norm": 0.054752644151449203,
+      "learning_rate": 0.00011127728189107576,
+      "loss": 0.7045,
+      "step": 1674
+    },
+    {
+      "epoch": 0.93,
+      "grad_norm": 0.0511183924973011,
+      "learning_rate": 0.00011119022928115007,
+      "loss": 0.6525,
+      "step": 1675
+    },
+    {
+      "epoch": 0.93,
+      "grad_norm": 0.05391978099942207,
+      "learning_rate": 0.00011110316808270831,
+      "loss": 0.6324,
+      "step": 1676
+    },
+    {
+      "epoch": 0.93,
+      "grad_norm": 0.05610180273652077,
+      "learning_rate": 0.00011101609836257008,
+      "loss": 0.6382,
+      "step": 1677
+    },
+    {
+      "epoch": 0.93,
+      "grad_norm": 0.05987925082445145,
+      "learning_rate": 0.00011092902018756151,
+      "loss": 0.7006,
+      "step": 1678
+    },
+    {
+      "epoch": 0.93,
+      "grad_norm": 0.06315992027521133,
+      "learning_rate": 0.0001108419336245152,
+      "loss": 0.7349,
+      "step": 1679
+    },
+    {
+      "epoch": 0.93,
+      "grad_norm": 0.05764647200703621,
+      "learning_rate": 0.0001107548387402702,
+      "loss": 0.6374,
+      "step": 1680
+    },
+    {
+      "epoch": 0.93,
+      "grad_norm": 0.05375726893544197,
+      "learning_rate": 0.00011066773560167196,
+      "loss": 0.6368,
+      "step": 1681
+    },
+    {
+      "epoch": 0.94,
+      "grad_norm": 0.05238979682326317,
+      "learning_rate": 0.00011058062427557229,
+      "loss": 0.5693,
+      "step": 1682
+    },
+    {
+      "epoch": 0.94,
+      "grad_norm": 0.055107586085796356,
+      "learning_rate": 0.00011049350482882919,
+      "loss": 0.6024,
+      "step": 1683
+    },
+    {
+      "epoch": 0.94,
+      "grad_norm": 0.0552542470395565,
+      "learning_rate": 0.00011040637732830701,
+      "loss": 0.5507,
+      "step": 1684
+    },
+    {
+      "epoch": 0.94,
+      "grad_norm": 0.06206995248794556,
+      "learning_rate": 0.00011031924184087618,
+      "loss": 0.7039,
+      "step": 1685
+    },
+    {
+      "epoch": 0.94,
+      "grad_norm": 0.05662161484360695,
+      "learning_rate": 0.00011023209843341333,
+      "loss": 0.6329,
+      "step": 1686
+    },
+    {
+      "epoch": 0.94,
+      "grad_norm": 0.060033902525901794,
+      "learning_rate": 0.00011014494717280115,
+      "loss": 0.6859,
+      "step": 1687
+    },
+    {
+      "epoch": 0.94,
+      "grad_norm": 0.05562319979071617,
+      "learning_rate": 0.00011005778812592832,
+      "loss": 0.6442,
+      "step": 1688
+    },
+    {
+      "epoch": 0.94,
+      "grad_norm": 0.05592164769768715,
+      "learning_rate": 0.00010997062135968956,
+      "loss": 0.6388,
+      "step": 1689
+    },
+    {
+      "epoch": 0.94,
+      "grad_norm": 0.05650092288851738,
+      "learning_rate": 0.00010988344694098545,
+      "loss": 0.6502,
+      "step": 1690
+    },
+    {
+      "epoch": 0.94,
+      "grad_norm": 0.06203974410891533,
+      "learning_rate": 0.00010979626493672245,
+      "loss": 0.7274,
+      "step": 1691
+    },
+    {
+      "epoch": 0.94,
+      "grad_norm": 0.052279599010944366,
+      "learning_rate": 0.00010970907541381295,
+      "loss": 0.5828,
+      "step": 1692
+    },
+    {
+      "epoch": 0.94,
+      "grad_norm": 0.06009140610694885,
+      "learning_rate": 0.00010962187843917497,
+      "loss": 0.6759,
+      "step": 1693
+    },
+    {
+      "epoch": 0.94,
+      "grad_norm": 0.05527809262275696,
+      "learning_rate": 0.0001095346740797323,
+      "loss": 0.6657,
+      "step": 1694
+    },
+    {
+      "epoch": 0.94,
+      "grad_norm": 0.0608220137655735,
+      "learning_rate": 0.00010944746240241444,
+      "loss": 0.6928,
+      "step": 1695
+    },
+    {
+      "epoch": 0.94,
+      "grad_norm": 0.05583418905735016,
+      "learning_rate": 0.00010936024347415643,
+      "loss": 0.5839,
+      "step": 1696
+    },
+    {
+      "epoch": 0.94,
+      "grad_norm": 0.05640941858291626,
+      "learning_rate": 0.00010927301736189893,
+      "loss": 0.661,
+      "step": 1697
+    },
+    {
+      "epoch": 0.94,
+      "grad_norm": 0.0547964982688427,
+      "learning_rate": 0.0001091857841325881,
+      "loss": 0.562,
+      "step": 1698
+    },
+    {
+      "epoch": 0.94,
+      "grad_norm": 0.05608777329325676,
+      "learning_rate": 0.00010909854385317557,
+      "loss": 0.5974,
+      "step": 1699
+    },
+    {
+      "epoch": 0.95,
+      "grad_norm": 0.05370324105024338,
+      "learning_rate": 0.00010901129659061837,
+      "loss": 0.6899,
+      "step": 1700
+    },
+    {
+      "epoch": 0.95,
+      "grad_norm": 0.06166364252567291,
+      "learning_rate": 0.00010892404241187886,
+      "loss": 0.6538,
+      "step": 1701
+    },
+    {
+      "epoch": 0.95,
+      "grad_norm": 0.0633421465754509,
+      "learning_rate": 0.00010883678138392477,
+      "loss": 0.785,
+      "step": 1702
+    },
+    {
+      "epoch": 0.95,
+      "grad_norm": 0.06182889640331268,
+      "learning_rate": 0.00010874951357372906,
+      "loss": 0.6403,
+      "step": 1703
+    },
+    {
+      "epoch": 0.95,
+      "grad_norm": 0.06533250212669373,
+      "learning_rate": 0.0001086622390482699,
+      "loss": 0.7551,
+      "step": 1704
+    },
+    {
+      "epoch": 0.95,
+      "grad_norm": 0.055042069405317307,
+      "learning_rate": 0.00010857495787453058,
+      "loss": 0.6083,
+      "step": 1705
+    },
+    {
+      "epoch": 0.95,
+      "grad_norm": 0.06447125971317291,
+      "learning_rate": 0.00010848767011949952,
+      "loss": 0.615,
+      "step": 1706
+    },
+    {
+      "epoch": 0.95,
+      "grad_norm": 0.056245360523462296,
+      "learning_rate": 0.00010840037585017022,
+      "loss": 0.6705,
+      "step": 1707
+    },
+    {
+      "epoch": 0.95,
+      "grad_norm": 0.05670240893959999,
+      "learning_rate": 0.00010831307513354112,
+      "loss": 0.6491,
+      "step": 1708
+    },
+    {
+      "epoch": 0.95,
+      "grad_norm": 0.05644702538847923,
+      "learning_rate": 0.00010822576803661564,
+      "loss": 0.6462,
+      "step": 1709
+    },
+    {
+      "epoch": 0.95,
+      "grad_norm": 0.05394309014081955,
+      "learning_rate": 0.00010813845462640206,
+      "loss": 0.6643,
+      "step": 1710
+    },
+    {
+      "epoch": 0.95,
+      "grad_norm": 0.057082321494817734,
+      "learning_rate": 0.00010805113496991364,
+      "loss": 0.7565,
+      "step": 1711
+    },
+    {
+      "epoch": 0.95,
+      "grad_norm": 0.06196596845984459,
+      "learning_rate": 0.00010796380913416823,
+      "loss": 0.6655,
+      "step": 1712
+    },
+    {
+      "epoch": 0.95,
+      "grad_norm": 0.06284771114587784,
+      "learning_rate": 0.0001078764771861886,
+      "loss": 0.647,
+      "step": 1713
+    },
+    {
+      "epoch": 0.95,
+      "grad_norm": 0.06070135906338692,
+      "learning_rate": 0.0001077891391930021,
+      "loss": 0.6143,
+      "step": 1714
+    },
+    {
+      "epoch": 0.95,
+      "grad_norm": 0.058664754033088684,
+      "learning_rate": 0.00010770179522164079,
+      "loss": 0.6618,
+      "step": 1715
+    },
+    {
+      "epoch": 0.95,
+      "grad_norm": 0.0531286895275116,
+      "learning_rate": 0.00010761444533914125,
+      "loss": 0.6304,
+      "step": 1716
+    },
+    {
+      "epoch": 0.95,
+      "grad_norm": 0.053345970809459686,
+      "learning_rate": 0.0001075270896125446,
+      "loss": 0.6148,
+      "step": 1717
+    },
+    {
+      "epoch": 0.96,
+      "grad_norm": 0.05614785850048065,
+      "learning_rate": 0.00010743972810889655,
+      "loss": 0.5945,
+      "step": 1718
+    },
+    {
+      "epoch": 0.96,
+      "grad_norm": 0.05940824747085571,
+      "learning_rate": 0.00010735236089524716,
+      "loss": 0.6681,
+      "step": 1719
+    },
+    {
+      "epoch": 0.96,
+      "grad_norm": 0.0561952069401741,
+      "learning_rate": 0.00010726498803865088,
+      "loss": 0.6925,
+      "step": 1720
+    },
+    {
+      "epoch": 0.96,
+      "grad_norm": 0.053867727518081665,
+      "learning_rate": 0.00010717760960616643,
+      "loss": 0.5798,
+      "step": 1721
+    },
+    {
+      "epoch": 0.96,
+      "grad_norm": 0.06327812373638153,
+      "learning_rate": 0.00010709022566485698,
+      "loss": 0.7174,
+      "step": 1722
+    },
+    {
+      "epoch": 0.96,
+      "grad_norm": 0.06408220529556274,
+      "learning_rate": 0.00010700283628178975,
+      "loss": 0.7552,
+      "step": 1723
+    },
+    {
+      "epoch": 0.96,
+      "grad_norm": 0.06066514551639557,
+      "learning_rate": 0.00010691544152403623,
+      "loss": 0.72,
+      "step": 1724
+    },
+    {
+      "epoch": 0.96,
+      "grad_norm": 0.06779921799898148,
+      "learning_rate": 0.00010682804145867204,
+      "loss": 0.7299,
+      "step": 1725
+    },
+    {
+      "epoch": 0.96,
+      "grad_norm": 0.060735609382390976,
+      "learning_rate": 0.0001067406361527768,
+      "loss": 0.6431,
+      "step": 1726
+    },
+    {
+      "epoch": 0.96,
+      "grad_norm": 0.05965111032128334,
+      "learning_rate": 0.00010665322567343423,
+      "loss": 0.6426,
+      "step": 1727
+    },
+    {
+      "epoch": 0.96,
+      "grad_norm": 0.05794420465826988,
+      "learning_rate": 0.00010656581008773198,
+      "loss": 0.6485,
+      "step": 1728
+    },
+    {
+      "epoch": 0.96,
+      "grad_norm": 0.06031234934926033,
+      "learning_rate": 0.00010647838946276165,
+      "loss": 0.6548,
+      "step": 1729
+    },
+    {
+      "epoch": 0.96,
+      "grad_norm": 0.056517090648412704,
+      "learning_rate": 0.00010639096386561864,
+      "loss": 0.6871,
+      "step": 1730
+    },
+    {
+      "epoch": 0.96,
+      "grad_norm": 0.06209394708275795,
+      "learning_rate": 0.00010630353336340226,
+      "loss": 0.6645,
+      "step": 1731
+    },
+    {
+      "epoch": 0.96,
+      "grad_norm": 0.059892479330301285,
+      "learning_rate": 0.00010621609802321555,
+      "loss": 0.6261,
+      "step": 1732
+    },
+    {
+      "epoch": 0.96,
+      "grad_norm": 0.06408750265836716,
+      "learning_rate": 0.0001061286579121652,
+      "loss": 0.7269,
+      "step": 1733
+    },
+    {
+      "epoch": 0.96,
+      "grad_norm": 0.057373423129320145,
+      "learning_rate": 0.00010604121309736164,
+      "loss": 0.6173,
+      "step": 1734
+    },
+    {
+      "epoch": 0.96,
+      "grad_norm": 0.055811841040849686,
+      "learning_rate": 0.00010595376364591889,
+      "loss": 0.6444,
+      "step": 1735
+    },
+    {
+      "epoch": 0.97,
+      "grad_norm": 0.05903814360499382,
+      "learning_rate": 0.00010586630962495452,
+      "loss": 0.7005,
+      "step": 1736
+    },
+    {
+      "epoch": 0.97,
+      "grad_norm": 0.06342557072639465,
+      "learning_rate": 0.00010577885110158958,
+      "loss": 0.6226,
+      "step": 1737
+    },
+    {
+      "epoch": 0.97,
+      "grad_norm": 0.06033981963992119,
+      "learning_rate": 0.00010569138814294864,
+      "loss": 0.6606,
+      "step": 1738
+    },
+    {
+      "epoch": 0.97,
+      "grad_norm": 0.0658353865146637,
+      "learning_rate": 0.00010560392081615962,
+      "loss": 0.6379,
+      "step": 1739
+    },
+    {
+      "epoch": 0.97,
+      "grad_norm": 0.05433519929647446,
+      "learning_rate": 0.00010551644918835381,
+      "loss": 0.5832,
+      "step": 1740
+    },
+    {
+      "epoch": 0.97,
+      "grad_norm": 0.058659233152866364,
+      "learning_rate": 0.00010542897332666581,
+      "loss": 0.6252,
+      "step": 1741
+    },
+    {
+      "epoch": 0.97,
+      "grad_norm": 0.06791430711746216,
+      "learning_rate": 0.00010534149329823349,
+      "loss": 0.7116,
+      "step": 1742
+    },
+    {
+      "epoch": 0.97,
+      "grad_norm": 0.06629349291324615,
+      "learning_rate": 0.00010525400917019784,
+      "loss": 0.9018,
+      "step": 1743
+    },
+    {
+      "epoch": 0.97,
+      "grad_norm": 0.05723525211215019,
+      "learning_rate": 0.00010516652100970308,
+      "loss": 0.6521,
+      "step": 1744
+    },
+    {
+      "epoch": 0.97,
+      "grad_norm": 0.0574793666601181,
+      "learning_rate": 0.00010507902888389647,
+      "loss": 0.65,
+      "step": 1745
+    },
+    {
+      "epoch": 0.97,
+      "grad_norm": 0.05418836697936058,
+      "learning_rate": 0.00010499153285992833,
+      "loss": 0.5964,
+      "step": 1746
+    },
+    {
+      "epoch": 0.97,
+      "grad_norm": 0.05051202327013016,
+      "learning_rate": 0.00010490403300495201,
+      "loss": 0.6072,
+      "step": 1747
+    },
+    {
+      "epoch": 0.97,
+      "grad_norm": 0.05381901189684868,
+      "learning_rate": 0.00010481652938612374,
+      "loss": 0.5846,
+      "step": 1748
+    },
+    {
+      "epoch": 0.97,
+      "grad_norm": 0.055063627660274506,
+      "learning_rate": 0.00010472902207060265,
+      "loss": 0.5894,
+      "step": 1749
+    },
+    {
+      "epoch": 0.97,
+      "grad_norm": 0.061404235661029816,
+      "learning_rate": 0.00010464151112555077,
+      "loss": 0.6717,
+      "step": 1750
+    },
+    {
+      "epoch": 0.97,
+      "grad_norm": 0.05025755986571312,
+      "learning_rate": 0.00010455399661813283,
+      "loss": 0.5984,
+      "step": 1751
+    },
+    {
+      "epoch": 0.97,
+      "grad_norm": 0.0560632087290287,
+      "learning_rate": 0.00010446647861551633,
+      "loss": 0.5581,
+      "step": 1752
+    },
+    {
+      "epoch": 0.97,
+      "grad_norm": 0.06142037361860275,
+      "learning_rate": 0.0001043789571848715,
+      "loss": 0.5679,
+      "step": 1753
+    },
+    {
+      "epoch": 0.98,
+      "grad_norm": 0.056344058364629745,
+      "learning_rate": 0.00010429143239337112,
+      "loss": 0.6612,
+      "step": 1754
+    },
+    {
+      "epoch": 0.98,
+      "grad_norm": 0.06066104397177696,
+      "learning_rate": 0.00010420390430819058,
+      "loss": 0.7477,
+      "step": 1755
+    },
+    {
+      "epoch": 0.98,
+      "grad_norm": 0.05510355904698372,
+      "learning_rate": 0.00010411637299650783,
+      "loss": 0.6163,
+      "step": 1756
+    },
+    {
+      "epoch": 0.98,
+      "grad_norm": 0.057294245809316635,
+      "learning_rate": 0.00010402883852550325,
+      "loss": 0.6468,
+      "step": 1757
+    },
+    {
+      "epoch": 0.98,
+      "grad_norm": 0.056174419820308685,
+      "learning_rate": 0.00010394130096235966,
+      "loss": 0.6213,
+      "step": 1758
+    },
+    {
+      "epoch": 0.98,
+      "grad_norm": 0.06194084510207176,
+      "learning_rate": 0.00010385376037426226,
+      "loss": 0.7077,
+      "step": 1759
+    },
+    {
+      "epoch": 0.98,
+      "grad_norm": 0.06177836284041405,
+      "learning_rate": 0.00010376621682839857,
+      "loss": 0.6536,
+      "step": 1760
+    },
+    {
+      "epoch": 0.98,
+      "grad_norm": 0.059714220464229584,
+      "learning_rate": 0.00010367867039195842,
+      "loss": 0.5912,
+      "step": 1761
+    },
+    {
+      "epoch": 0.98,
+      "grad_norm": 0.052113138139247894,
+      "learning_rate": 0.00010359112113213376,
+      "loss": 0.5401,
+      "step": 1762
+    },
+    {
+      "epoch": 0.98,
+      "grad_norm": 0.06121697649359703,
+      "learning_rate": 0.0001035035691161188,
+      "loss": 0.6862,
+      "step": 1763
+    },
+    {
+      "epoch": 0.98,
+      "grad_norm": 0.05976003035902977,
+      "learning_rate": 0.00010341601441110983,
+      "loss": 0.58,
+      "step": 1764
+    },
+    {
+      "epoch": 0.98,
+      "grad_norm": 0.05741668865084648,
+      "learning_rate": 0.00010332845708430519,
+      "loss": 0.5233,
+      "step": 1765
+    },
+    {
+      "epoch": 0.98,
+      "grad_norm": 0.06236950680613518,
+      "learning_rate": 0.00010324089720290521,
+      "loss": 0.7257,
+      "step": 1766
+    },
+    {
+      "epoch": 0.98,
+      "grad_norm": 0.06206139549612999,
+      "learning_rate": 0.00010315333483411232,
+      "loss": 0.6985,
+      "step": 1767
+    },
+    {
+      "epoch": 0.98,
+      "grad_norm": 0.05406549200415611,
+      "learning_rate": 0.00010306577004513065,
+      "loss": 0.5677,
+      "step": 1768
+    },
+    {
+      "epoch": 0.98,
+      "grad_norm": 0.06077892333269119,
+      "learning_rate": 0.0001029782029031663,
+      "loss": 0.6745,
+      "step": 1769
+    },
+    {
+      "epoch": 0.98,
+      "grad_norm": 0.06338007748126984,
+      "learning_rate": 0.00010289063347542726,
+      "loss": 0.6612,
+      "step": 1770
+    },
+    {
+      "epoch": 0.98,
+      "grad_norm": 0.060978084802627563,
+      "learning_rate": 0.00010280306182912313,
+      "loss": 0.7625,
+      "step": 1771
+    },
+    {
+      "epoch": 0.99,
+      "grad_norm": 0.0552746020257473,
+      "learning_rate": 0.00010271548803146526,
+      "loss": 0.5827,
+      "step": 1772
+    },
+    {
+      "epoch": 0.99,
+      "grad_norm": 0.05428915098309517,
+      "learning_rate": 0.00010262791214966668,
+      "loss": 0.6786,
+      "step": 1773
+    },
+    {
+      "epoch": 0.99,
+      "grad_norm": 0.057721514254808426,
+      "learning_rate": 0.00010254033425094197,
+      "loss": 0.5914,
+      "step": 1774
+    },
+    {
+      "epoch": 0.99,
+      "grad_norm": 0.05720565468072891,
+      "learning_rate": 0.00010245275440250728,
+      "loss": 0.6013,
+      "step": 1775
+    },
+    {
+      "epoch": 0.99,
+      "grad_norm": 0.05914613604545593,
+      "learning_rate": 0.00010236517267158028,
+      "loss": 0.6439,
+      "step": 1776
+    },
+    {
+      "epoch": 0.99,
+      "grad_norm": 0.057543400675058365,
+      "learning_rate": 0.00010227758912538008,
+      "loss": 0.7292,
+      "step": 1777
+    },
+    {
+      "epoch": 0.99,
+      "grad_norm": 0.056838035583496094,
+      "learning_rate": 0.00010219000383112713,
+      "loss": 0.6506,
+      "step": 1778
+    },
+    {
+      "epoch": 0.99,
+      "grad_norm": 0.05806044489145279,
+      "learning_rate": 0.00010210241685604331,
+      "loss": 0.6114,
+      "step": 1779
+    },
+    {
+      "epoch": 0.99,
+      "grad_norm": 0.0533684603869915,
+      "learning_rate": 0.00010201482826735172,
+      "loss": 0.5946,
+      "step": 1780
+    },
+    {
+      "epoch": 0.99,
+      "grad_norm": 0.05922561138868332,
+      "learning_rate": 0.00010192723813227672,
+      "loss": 0.6626,
+      "step": 1781
+    },
+    {
+      "epoch": 0.99,
+      "grad_norm": 0.055819686502218246,
+      "learning_rate": 0.00010183964651804382,
+      "loss": 0.6422,
+      "step": 1782
+    },
+    {
+      "epoch": 0.99,
+      "grad_norm": 0.050924383103847504,
+      "learning_rate": 0.00010175205349187977,
+      "loss": 0.5344,
+      "step": 1783
+    },
+    {
+      "epoch": 0.99,
+      "grad_norm": 0.06296874582767487,
+      "learning_rate": 0.00010166445912101231,
+      "loss": 0.6451,
+      "step": 1784
+    },
+    {
+      "epoch": 0.99,
+      "grad_norm": 0.05765317752957344,
+      "learning_rate": 0.00010157686347267021,
+      "loss": 0.6403,
+      "step": 1785
+    },
+    {
+      "epoch": 0.99,
+      "grad_norm": 0.060042284429073334,
+      "learning_rate": 0.00010148926661408327,
+      "loss": 0.6919,
+      "step": 1786
+    },
+    {
+      "epoch": 0.99,
+      "grad_norm": 0.06098796799778938,
+      "learning_rate": 0.0001014016686124822,
+      "loss": 0.6371,
+      "step": 1787
+    },
+    {
+      "epoch": 0.99,
+      "grad_norm": 0.05685941502451897,
+      "learning_rate": 0.00010131406953509857,
+      "loss": 0.6399,
+      "step": 1788
+    },
+    {
+      "epoch": 0.99,
+      "grad_norm": 0.06483474373817444,
+      "learning_rate": 0.00010122646944916483,
+      "loss": 0.6915,
+      "step": 1789
+    },
+    {
+      "epoch": 1.0,
+      "grad_norm": 0.06274085491895676,
+      "learning_rate": 0.00010113886842191408,
+      "loss": 0.6212,
+      "step": 1790
+    },
+    {
+      "epoch": 1.0,
+      "grad_norm": 0.06120248883962631,
+      "learning_rate": 0.00010105126652058032,
+      "loss": 0.6859,
+      "step": 1791
+    },
+    {
+      "epoch": 1.0,
+      "grad_norm": 0.06124896928668022,
+      "learning_rate": 0.00010096366381239808,
+      "loss": 0.5996,
+      "step": 1792
+    },
+    {
+      "epoch": 1.0,
+      "grad_norm": 0.05330495536327362,
+      "learning_rate": 0.00010087606036460257,
+      "loss": 0.7189,
+      "step": 1793
+    },
+    {
+      "epoch": 1.0,
+      "grad_norm": 0.06091773882508278,
+      "learning_rate": 0.00010078845624442954,
+      "loss": 0.6692,
+      "step": 1794
+    },
+    {
+      "epoch": 1.0,
+      "grad_norm": 0.057791825383901596,
+      "learning_rate": 0.0001007008515191153,
+      "loss": 0.6403,
+      "step": 1795
+    },
+    {
+      "epoch": 1.0,
+      "grad_norm": 0.06233106181025505,
+      "learning_rate": 0.00010061324625589657,
+      "loss": 0.6166,
+      "step": 1796
+    },
+    {
+      "epoch": 1.0,
+      "grad_norm": 0.06440749019384384,
+      "learning_rate": 0.0001005256405220105,
+      "loss": 0.7032,
+      "step": 1797
+    },
+    {
+      "epoch": 1.0,
+      "grad_norm": 0.060089919716119766,
+      "learning_rate": 0.00010043803438469461,
+      "loss": 0.6484,
+      "step": 1798
+    }
+  ],
+  "logging_steps": 1,
+  "max_steps": 3596,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 2,
+  "save_steps": 1798,
+  "total_flos": 3.21854445811925e+17,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}