{
  "best_global_step": 42000,
  "best_metric": 0.9448174834251404,
  "best_model_checkpoint": "./speecht5/urdu-tts-phonemes-speecht5-finetuned/checkpoint-42000",
  "epoch": 18.308732148697263,
  "eval_steps": 1000,
  "global_step": 42000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.08721247138340783,
      "grad_norm": 24.05974769592285,
      "learning_rate": 1.9900000000000004e-06,
      "loss": 8.479815063476563,
      "step": 200
    },
    {
      "epoch": 0.17442494276681567,
      "grad_norm": 20.256372451782227,
      "learning_rate": 3.990000000000001e-06,
      "loss": 6.65532470703125,
      "step": 400
    },
    {
      "epoch": 0.26163741415022346,
      "grad_norm": 24.020288467407227,
      "learning_rate": 5.99e-06,
      "loss": 5.549420776367188,
      "step": 600
    },
    {
      "epoch": 0.34884988553363133,
      "grad_norm": 22.95574188232422,
      "learning_rate": 7.990000000000001e-06,
      "loss": 5.242913818359375,
      "step": 800
    },
    {
      "epoch": 0.43606235691703915,
      "grad_norm": 21.500507354736328,
      "learning_rate": 9.990000000000001e-06,
      "loss": 5.012727661132812,
      "step": 1000
    },
    {
      "epoch": 0.43606235691703915,
      "eval_loss": 1.120340347290039,
      "eval_runtime": 331.3693,
      "eval_samples_per_second": 36.91,
      "eval_steps_per_second": 9.228,
      "step": 1000
    },
    {
      "epoch": 0.5232748283004469,
      "grad_norm": 24.796592712402344,
      "learning_rate": 9.978074041427943e-06,
      "loss": 4.87976318359375,
      "step": 1200
    },
    {
      "epoch": 0.6104872996838548,
      "grad_norm": 30.302698135375977,
      "learning_rate": 9.956037902159543e-06,
      "loss": 4.807673034667968,
      "step": 1400
    },
    {
      "epoch": 0.6976997710672627,
      "grad_norm": 26.418174743652344,
      "learning_rate": 9.934001762891142e-06,
      "loss": 4.722581787109375,
      "step": 1600
    },
    {
      "epoch": 0.7849122424506705,
      "grad_norm": 18.969362258911133,
      "learning_rate": 9.911965623622742e-06,
      "loss": 4.6823681640625,
      "step": 1800
    },
    {
      "epoch": 0.8721247138340783,
      "grad_norm": 31.632402420043945,
      "learning_rate": 9.889929484354342e-06,
      "loss": 4.643733215332031,
      "step": 2000
    },
    {
      "epoch": 0.8721247138340783,
      "eval_loss": 1.0700525045394897,
      "eval_runtime": 349.2614,
      "eval_samples_per_second": 35.02,
      "eval_steps_per_second": 8.756,
      "step": 2000
    },
    {
      "epoch": 0.9593371852174861,
      "grad_norm": 26.039878845214844,
      "learning_rate": 9.86789334508594e-06,
      "loss": 4.633777465820312,
      "step": 2200
    },
    {
      "epoch": 1.0462226098332061,
      "grad_norm": 29.846647262573242,
      "learning_rate": 9.845857205817542e-06,
      "loss": 4.54459716796875,
      "step": 2400
    },
    {
      "epoch": 1.133435081216614,
      "grad_norm": 30.55405616760254,
      "learning_rate": 9.823821066549142e-06,
      "loss": 4.533556518554687,
      "step": 2600
    },
    {
      "epoch": 1.2206475526000218,
      "grad_norm": 36.22990417480469,
      "learning_rate": 9.801784927280741e-06,
      "loss": 4.509813842773437,
      "step": 2800
    },
    {
      "epoch": 1.3078600239834297,
      "grad_norm": 23.28966522216797,
      "learning_rate": 9.77974878801234e-06,
      "loss": 4.494954528808594,
      "step": 3000
    },
    {
      "epoch": 1.3078600239834297,
      "eval_loss": 1.0368597507476807,
      "eval_runtime": 340.3775,
      "eval_samples_per_second": 35.934,
      "eval_steps_per_second": 8.984,
      "step": 3000
    },
    {
      "epoch": 1.3950724953668374,
      "grad_norm": 36.652828216552734,
      "learning_rate": 9.757712648743941e-06,
      "loss": 4.494131164550781,
      "step": 3200
    },
    {
      "epoch": 1.4822849667502453,
      "grad_norm": 21.791095733642578,
      "learning_rate": 9.73567650947554e-06,
      "loss": 4.468922119140625,
      "step": 3400
    },
    {
      "epoch": 1.569497438133653,
      "grad_norm": 36.221858978271484,
      "learning_rate": 9.71364037020714e-06,
      "loss": 4.425529479980469,
      "step": 3600
    },
    {
      "epoch": 1.656709909517061,
      "grad_norm": 25.282238006591797,
      "learning_rate": 9.69160423093874e-06,
      "loss": 4.403593139648438,
      "step": 3800
    },
    {
      "epoch": 1.743922380900469,
      "grad_norm": 21.640579223632812,
      "learning_rate": 9.669568091670341e-06,
      "loss": 4.409559326171875,
      "step": 4000
    },
    {
      "epoch": 1.743922380900469,
      "eval_loss": 1.0254037380218506,
      "eval_runtime": 353.8731,
      "eval_samples_per_second": 34.563,
      "eval_steps_per_second": 8.642,
      "step": 4000
    },
    {
      "epoch": 1.8311348522838766,
      "grad_norm": 34.924861907958984,
      "learning_rate": 9.64753195240194e-06,
      "loss": 4.403922119140625,
      "step": 4200
    },
    {
      "epoch": 1.9183473236672843,
      "grad_norm": 26.611953735351562,
      "learning_rate": 9.62549581313354e-06,
      "loss": 4.3958837890625,
      "step": 4400
    },
    {
      "epoch": 2.0052327482830044,
      "grad_norm": 24.899072647094727,
      "learning_rate": 9.60345967386514e-06,
      "loss": 4.356022338867188,
      "step": 4600
    },
    {
      "epoch": 2.0924452196664123,
      "grad_norm": 26.694162368774414,
      "learning_rate": 9.58142353459674e-06,
      "loss": 4.358612060546875,
      "step": 4800
    },
    {
      "epoch": 2.17965769104982,
      "grad_norm": 19.83888816833496,
      "learning_rate": 9.559387395328338e-06,
      "loss": 4.367332458496094,
      "step": 5000
    },
    {
      "epoch": 2.17965769104982,
      "eval_loss": 1.0280553102493286,
      "eval_runtime": 368.049,
      "eval_samples_per_second": 33.232,
      "eval_steps_per_second": 8.309,
      "step": 5000
    },
    {
      "epoch": 2.266870162433228,
      "grad_norm": 25.698699951171875,
      "learning_rate": 9.53735125605994e-06,
      "loss": 4.343710327148438,
      "step": 5200
    },
    {
      "epoch": 2.3540826338166356,
      "grad_norm": 21.149288177490234,
      "learning_rate": 9.51531511679154e-06,
      "loss": 4.331655883789063,
      "step": 5400
    },
    {
      "epoch": 2.4412951052000436,
      "grad_norm": 22.230365753173828,
      "learning_rate": 9.493278977523137e-06,
      "loss": 4.33089599609375,
      "step": 5600
    },
    {
      "epoch": 2.5285075765834515,
      "grad_norm": 38.653480529785156,
      "learning_rate": 9.471242838254738e-06,
      "loss": 4.331998291015625,
      "step": 5800
    },
    {
      "epoch": 2.6157200479668594,
      "grad_norm": 22.251720428466797,
      "learning_rate": 9.449206698986339e-06,
      "loss": 4.327017211914063,
      "step": 6000
    },
    {
      "epoch": 2.6157200479668594,
      "eval_loss": 1.008265495300293,
      "eval_runtime": 346.3793,
      "eval_samples_per_second": 35.311,
      "eval_steps_per_second": 8.828,
      "step": 6000
    },
    {
      "epoch": 2.7029325193502673,
      "grad_norm": 29.327293395996094,
      "learning_rate": 9.427170559717938e-06,
      "loss": 4.314130859375,
      "step": 6200
    },
    {
      "epoch": 2.790144990733675,
      "grad_norm": 42.60088348388672,
      "learning_rate": 9.405134420449537e-06,
      "loss": 4.301522827148437,
      "step": 6400
    },
    {
      "epoch": 2.8773574621170828,
      "grad_norm": 22.110036849975586,
      "learning_rate": 9.383098281181138e-06,
      "loss": 4.276791687011719,
      "step": 6600
    },
    {
      "epoch": 2.9645699335004907,
      "grad_norm": 25.4754581451416,
      "learning_rate": 9.361062141912737e-06,
      "loss": 4.284074096679688,
      "step": 6800
    },
    {
      "epoch": 3.0514553581162107,
      "grad_norm": 24.357316970825195,
      "learning_rate": 9.339026002644338e-06,
      "loss": 4.260386962890625,
      "step": 7000
    },
    {
      "epoch": 3.0514553581162107,
      "eval_loss": 1.0023632049560547,
      "eval_runtime": 346.3632,
      "eval_samples_per_second": 35.313,
      "eval_steps_per_second": 8.829,
      "step": 7000
    },
    {
      "epoch": 3.1386678294996186,
      "grad_norm": 25.026887893676758,
      "learning_rate": 9.316989863375937e-06,
      "loss": 4.281555786132812,
      "step": 7200
    },
    {
      "epoch": 3.225880300883026,
      "grad_norm": 24.85771942138672,
      "learning_rate": 9.294953724107538e-06,
      "loss": 4.265211181640625,
      "step": 7400
    },
    {
      "epoch": 3.313092772266434,
      "grad_norm": 29.351341247558594,
      "learning_rate": 9.272917584839137e-06,
      "loss": 4.274870300292969,
      "step": 7600
    },
    {
      "epoch": 3.400305243649842,
      "grad_norm": 29.364891052246094,
      "learning_rate": 9.250881445570736e-06,
      "loss": 4.261960144042969,
      "step": 7800
    },
    {
      "epoch": 3.48751771503325,
      "grad_norm": 22.959003448486328,
      "learning_rate": 9.228845306302337e-06,
      "loss": 4.241669921875,
      "step": 8000
    },
    {
      "epoch": 3.48751771503325,
      "eval_loss": 0.992875874042511,
      "eval_runtime": 349.2513,
      "eval_samples_per_second": 35.021,
      "eval_steps_per_second": 8.756,
      "step": 8000
    },
    {
      "epoch": 3.574730186416658,
      "grad_norm": 24.052427291870117,
      "learning_rate": 9.206809167033936e-06,
      "loss": 4.246179504394531,
      "step": 8200
    },
    {
      "epoch": 3.6619426578000653,
      "grad_norm": 23.832117080688477,
      "learning_rate": 9.184773027765537e-06,
      "loss": 4.259029541015625,
      "step": 8400
    },
    {
      "epoch": 3.7491551291834733,
      "grad_norm": 18.504886627197266,
      "learning_rate": 9.162736888497136e-06,
      "loss": 4.235401916503906,
      "step": 8600
    },
    {
      "epoch": 3.836367600566881,
      "grad_norm": 30.796070098876953,
      "learning_rate": 9.140700749228737e-06,
      "loss": 4.245496826171875,
      "step": 8800
    },
    {
      "epoch": 3.9235800719502887,
      "grad_norm": 39.91634750366211,
      "learning_rate": 9.118664609960336e-06,
      "loss": 4.229890747070312,
      "step": 9000
    },
    {
      "epoch": 3.9235800719502887,
      "eval_loss": 0.9932724833488464,
      "eval_runtime": 346.2201,
      "eval_samples_per_second": 35.327,
      "eval_steps_per_second": 8.833,
      "step": 9000
    },
    {
      "epoch": 4.010465496566009,
      "grad_norm": 26.46392822265625,
      "learning_rate": 9.096628470691935e-06,
      "loss": 4.214105224609375,
      "step": 9200
    },
    {
      "epoch": 4.097677967949417,
      "grad_norm": 43.10508728027344,
      "learning_rate": 9.074592331423536e-06,
      "loss": 4.208152465820312,
      "step": 9400
    },
    {
      "epoch": 4.184890439332825,
      "grad_norm": 35.05559158325195,
      "learning_rate": 9.052556192155135e-06,
      "loss": 4.218372802734375,
      "step": 9600
    },
    {
      "epoch": 4.2721029107162325,
      "grad_norm": 30.729036331176758,
      "learning_rate": 9.030520052886736e-06,
      "loss": 4.209480590820313,
      "step": 9800
    },
    {
      "epoch": 4.35931538209964,
      "grad_norm": 31.454877853393555,
      "learning_rate": 9.008483913618335e-06,
      "loss": 4.210875854492188,
      "step": 10000
    },
    {
      "epoch": 4.35931538209964,
      "eval_loss": 0.992638349533081,
      "eval_runtime": 352.192,
      "eval_samples_per_second": 34.728,
      "eval_steps_per_second": 8.683,
      "step": 10000
    },
    {
      "epoch": 4.446527853483048,
      "grad_norm": 28.604127883911133,
      "learning_rate": 8.986447774349934e-06,
      "loss": 4.210906982421875,
      "step": 10200
    },
    {
      "epoch": 4.533740324866456,
      "grad_norm": 23.0704345703125,
      "learning_rate": 8.964411635081535e-06,
      "loss": 4.199930419921875,
      "step": 10400
    },
    {
      "epoch": 4.620952796249863,
      "grad_norm": 26.80572509765625,
      "learning_rate": 8.942375495813135e-06,
      "loss": 4.209544677734375,
      "step": 10600
    },
    {
      "epoch": 4.708165267633271,
      "grad_norm": 22.299110412597656,
      "learning_rate": 8.920339356544735e-06,
      "loss": 4.200810546875,
      "step": 10800
    },
    {
      "epoch": 4.795377739016679,
      "grad_norm": 22.755102157592773,
      "learning_rate": 8.898303217276334e-06,
      "loss": 4.196201477050781,
      "step": 11000
    },
    {
      "epoch": 4.795377739016679,
      "eval_loss": 0.987343430519104,
      "eval_runtime": 352.3769,
      "eval_samples_per_second": 34.71,
      "eval_steps_per_second": 8.678,
      "step": 11000
    },
    {
      "epoch": 4.882590210400087,
      "grad_norm": 22.76228904724121,
      "learning_rate": 8.876267078007934e-06,
      "loss": 4.192545776367187,
      "step": 11200
    },
    {
      "epoch": 4.969802681783495,
      "grad_norm": 18.565929412841797,
      "learning_rate": 8.854230938739534e-06,
      "loss": 4.180068359375,
      "step": 11400
    },
    {
      "epoch": 5.056688106399215,
      "grad_norm": 21.868263244628906,
      "learning_rate": 8.832194799471133e-06,
      "loss": 4.163085327148438,
      "step": 11600
    },
    {
      "epoch": 5.143900577782623,
      "grad_norm": 31.216304779052734,
      "learning_rate": 8.810158660202733e-06,
      "loss": 4.174065246582031,
      "step": 11800
    },
    {
      "epoch": 5.231113049166031,
      "grad_norm": 24.363231658935547,
      "learning_rate": 8.788122520934334e-06,
      "loss": 4.187493896484375,
      "step": 12000
    },
    {
      "epoch": 5.231113049166031,
      "eval_loss": 0.9810986518859863,
      "eval_runtime": 339.4129,
      "eval_samples_per_second": 36.036,
      "eval_steps_per_second": 9.01,
      "step": 12000
    },
    {
      "epoch": 5.318325520549439,
      "grad_norm": 19.92061424255371,
      "learning_rate": 8.766086381665933e-06,
      "loss": 4.187520141601563,
      "step": 12200
    },
    {
      "epoch": 5.405537991932847,
      "grad_norm": 24.99143409729004,
      "learning_rate": 8.744050242397532e-06,
      "loss": 4.166245727539063,
      "step": 12400
    },
    {
      "epoch": 5.492750463316254,
      "grad_norm": 24.70585823059082,
      "learning_rate": 8.722014103129133e-06,
      "loss": 4.164942016601563,
      "step": 12600
    },
    {
      "epoch": 5.579962934699662,
      "grad_norm": 18.944747924804688,
      "learning_rate": 8.699977963860732e-06,
      "loss": 4.16548095703125,
      "step": 12800
    },
    {
      "epoch": 5.66717540608307,
      "grad_norm": 24.14280891418457,
      "learning_rate": 8.677941824592331e-06,
      "loss": 4.168301391601562,
      "step": 13000
    },
    {
      "epoch": 5.66717540608307,
      "eval_loss": 0.9759051203727722,
      "eval_runtime": 349.8176,
      "eval_samples_per_second": 34.964,
      "eval_steps_per_second": 8.742,
      "step": 13000
    },
    {
      "epoch": 5.754387877466478,
      "grad_norm": 23.709392547607422,
      "learning_rate": 8.655905685323932e-06,
      "loss": 4.159636840820313,
      "step": 13200
    },
    {
      "epoch": 5.8416003488498855,
      "grad_norm": 37.89950942993164,
      "learning_rate": 8.633869546055531e-06,
      "loss": 4.171852416992188,
      "step": 13400
    },
    {
      "epoch": 5.9288128202332935,
      "grad_norm": 34.13504409790039,
      "learning_rate": 8.61183340678713e-06,
      "loss": 4.164423828125,
      "step": 13600
    },
    {
      "epoch": 6.0156982448490135,
      "grad_norm": 24.551559448242188,
      "learning_rate": 8.589797267518731e-06,
      "loss": 4.130006103515625,
      "step": 13800
    },
    {
      "epoch": 6.102910716232421,
      "grad_norm": 21.308082580566406,
      "learning_rate": 8.567761128250332e-06,
      "loss": 4.149431762695312,
      "step": 14000
    },
    {
      "epoch": 6.102910716232421,
      "eval_loss": 0.9769551157951355,
      "eval_runtime": 342.9942,
      "eval_samples_per_second": 35.659,
      "eval_steps_per_second": 8.916,
      "step": 14000
    },
    {
      "epoch": 6.190123187615829,
      "grad_norm": 22.955293655395508,
      "learning_rate": 8.545724988981931e-06,
      "loss": 4.142289428710938,
      "step": 14200
    },
    {
      "epoch": 6.277335658999237,
      "grad_norm": 17.721412658691406,
      "learning_rate": 8.52368884971353e-06,
      "loss": 4.137732238769531,
      "step": 14400
    },
    {
      "epoch": 6.364548130382644,
      "grad_norm": 25.668283462524414,
      "learning_rate": 8.501652710445131e-06,
      "loss": 4.1462646484375,
      "step": 14600
    },
    {
      "epoch": 6.451760601766052,
      "grad_norm": 28.969520568847656,
      "learning_rate": 8.47961657117673e-06,
      "loss": 4.1304473876953125,
      "step": 14800
    },
    {
      "epoch": 6.53897307314946,
      "grad_norm": 21.27937889099121,
      "learning_rate": 8.45758043190833e-06,
      "loss": 4.140865173339844,
      "step": 15000
    },
    {
      "epoch": 6.53897307314946,
      "eval_loss": 0.9721247553825378,
      "eval_runtime": 353.3602,
      "eval_samples_per_second": 34.613,
      "eval_steps_per_second": 8.654,
      "step": 15000
    },
    {
      "epoch": 6.626185544532868,
      "grad_norm": 25.535720825195312,
      "learning_rate": 8.43554429263993e-06,
      "loss": 4.151261291503906,
      "step": 15200
    },
    {
      "epoch": 6.713398015916276,
      "grad_norm": 31.536725997924805,
      "learning_rate": 8.413508153371531e-06,
      "loss": 4.1409674072265625,
      "step": 15400
    },
    {
      "epoch": 6.800610487299684,
      "grad_norm": 21.910253524780273,
      "learning_rate": 8.39147201410313e-06,
      "loss": 4.137415161132813,
      "step": 15600
    },
    {
      "epoch": 6.887822958683092,
      "grad_norm": 25.541889190673828,
      "learning_rate": 8.369435874834729e-06,
      "loss": 4.14487548828125,
      "step": 15800
    },
    {
      "epoch": 6.9750354300665,
      "grad_norm": 28.3773136138916,
      "learning_rate": 8.34739973556633e-06,
      "loss": 4.114954833984375,
      "step": 16000
    },
    {
      "epoch": 6.9750354300665,
      "eval_loss": 0.9710648059844971,
      "eval_runtime": 346.4652,
      "eval_samples_per_second": 35.302,
      "eval_steps_per_second": 8.826,
      "step": 16000
    },
    {
      "epoch": 7.06192085468222,
      "grad_norm": 28.948652267456055,
      "learning_rate": 8.325363596297929e-06,
      "loss": 4.1155767822265625,
      "step": 16200
    },
    {
      "epoch": 7.149133326065628,
      "grad_norm": 29.551963806152344,
      "learning_rate": 8.303327457029528e-06,
      "loss": 4.115559387207031,
      "step": 16400
    },
    {
      "epoch": 7.236345797449035,
      "grad_norm": 20.035167694091797,
      "learning_rate": 8.281291317761129e-06,
      "loss": 4.10067138671875,
      "step": 16600
    },
    {
      "epoch": 7.323558268832443,
      "grad_norm": 28.89103126525879,
      "learning_rate": 8.259255178492728e-06,
      "loss": 4.108811645507813,
      "step": 16800
    },
    {
      "epoch": 7.410770740215851,
      "grad_norm": 33.26117706298828,
      "learning_rate": 8.237219039224329e-06,
      "loss": 4.110437622070313,
      "step": 17000
    },
    {
      "epoch": 7.410770740215851,
      "eval_loss": 0.9670585989952087,
      "eval_runtime": 380.6236,
      "eval_samples_per_second": 32.134,
      "eval_steps_per_second": 8.034,
      "step": 17000
    },
    {
      "epoch": 7.497983211599259,
      "grad_norm": 22.1575984954834,
      "learning_rate": 8.215182899955928e-06,
      "loss": 4.118037719726562,
      "step": 17200
    },
    {
      "epoch": 7.5851956829826666,
      "grad_norm": 34.508846282958984,
      "learning_rate": 8.193146760687529e-06,
      "loss": 4.103672790527344,
      "step": 17400
    },
    {
      "epoch": 7.6724081543660745,
      "grad_norm": 27.750104904174805,
      "learning_rate": 8.171110621419128e-06,
      "loss": 4.096029052734375,
      "step": 17600
    },
    {
      "epoch": 7.759620625749482,
      "grad_norm": 23.576847076416016,
      "learning_rate": 8.149074482150729e-06,
      "loss": 4.114128112792969,
      "step": 17800
    },
    {
      "epoch": 7.84683309713289,
      "grad_norm": 18.866933822631836,
      "learning_rate": 8.127038342882328e-06,
      "loss": 4.109539184570313,
      "step": 18000
    },
    {
      "epoch": 7.84683309713289,
      "eval_loss": 0.9662941694259644,
      "eval_runtime": 366.7048,
      "eval_samples_per_second": 33.354,
      "eval_steps_per_second": 8.339,
      "step": 18000
    },
    {
      "epoch": 7.934045568516298,
      "grad_norm": 22.053308486938477,
      "learning_rate": 8.105002203613927e-06,
      "loss": 4.101531982421875,
      "step": 18200
    },
    {
      "epoch": 8.020930993132017,
      "grad_norm": 27.721097946166992,
      "learning_rate": 8.082966064345528e-06,
      "loss": 4.078556518554688,
      "step": 18400
    },
    {
      "epoch": 8.108143464515425,
      "grad_norm": 16.3195858001709,
      "learning_rate": 8.060929925077127e-06,
      "loss": 4.113186645507812,
      "step": 18600
    },
    {
      "epoch": 8.195355935898833,
      "grad_norm": 24.274024963378906,
      "learning_rate": 8.038893785808728e-06,
      "loss": 4.102982788085938,
      "step": 18800
    },
    {
      "epoch": 8.282568407282241,
      "grad_norm": 44.398868560791016,
      "learning_rate": 8.016857646540327e-06,
      "loss": 4.077982177734375,
      "step": 19000
    },
    {
      "epoch": 8.282568407282241,
      "eval_loss": 0.9664719104766846,
      "eval_runtime": 385.6838,
      "eval_samples_per_second": 31.713,
      "eval_steps_per_second": 7.929,
      "step": 19000
    },
    {
      "epoch": 8.36978087866565,
      "grad_norm": 28.94335174560547,
      "learning_rate": 7.994821507271928e-06,
      "loss": 4.09763671875,
      "step": 19200
    },
    {
      "epoch": 8.456993350049057,
      "grad_norm": 33.86271667480469,
      "learning_rate": 7.972785368003527e-06,
      "loss": 4.082224731445312,
      "step": 19400
    },
    {
      "epoch": 8.544205821432465,
      "grad_norm": 27.749034881591797,
      "learning_rate": 7.950749228735126e-06,
      "loss": 4.094083251953125,
      "step": 19600
    },
    {
      "epoch": 8.631418292815873,
      "grad_norm": 23.508079528808594,
      "learning_rate": 7.928713089466727e-06,
      "loss": 4.110226440429687,
      "step": 19800
    },
    {
      "epoch": 8.71863076419928,
      "grad_norm": 24.7147216796875,
      "learning_rate": 7.906676950198326e-06,
      "loss": 4.091410522460937,
      "step": 20000
    },
    {
      "epoch": 8.71863076419928,
      "eval_loss": 0.9642230868339539,
      "eval_runtime": 348.2116,
      "eval_samples_per_second": 35.125,
      "eval_steps_per_second": 8.782,
      "step": 20000
    },
    {
      "epoch": 8.805843235582689,
      "grad_norm": 26.83739471435547,
      "learning_rate": 7.884640810929925e-06,
      "loss": 4.072673645019531,
      "step": 20200
    },
    {
      "epoch": 8.893055706966097,
      "grad_norm": 23.345003128051758,
      "learning_rate": 7.862604671661526e-06,
      "loss": 4.081375732421875,
      "step": 20400
    },
    {
      "epoch": 8.980268178349505,
      "grad_norm": 22.3059024810791,
      "learning_rate": 7.840568532393126e-06,
      "loss": 4.080589599609375,
      "step": 20600
    },
    {
      "epoch": 9.067153602965224,
      "grad_norm": 28.262033462524414,
      "learning_rate": 7.818532393124725e-06,
      "loss": 4.071205444335938,
      "step": 20800
    },
    {
      "epoch": 9.154366074348632,
      "grad_norm": 17.770212173461914,
      "learning_rate": 7.796496253856325e-06,
      "loss": 4.075708618164063,
      "step": 21000
    },
    {
      "epoch": 9.154366074348632,
      "eval_loss": 0.9607056379318237,
      "eval_runtime": 345.0951,
      "eval_samples_per_second": 35.442,
      "eval_steps_per_second": 8.861,
      "step": 21000
    },
    {
      "epoch": 9.24157854573204,
      "grad_norm": 26.18967056274414,
      "learning_rate": 7.774460114587925e-06,
      "loss": 4.063737487792968,
      "step": 21200
    },
    {
      "epoch": 9.328791017115448,
      "grad_norm": 20.26483154296875,
      "learning_rate": 7.752423975319524e-06,
      "loss": 4.061445922851562,
      "step": 21400
    },
    {
      "epoch": 9.416003488498855,
      "grad_norm": 36.16294479370117,
      "learning_rate": 7.730387836051124e-06,
      "loss": 4.046877136230469,
      "step": 21600
    },
    {
      "epoch": 9.503215959882263,
      "grad_norm": 31.608964920043945,
      "learning_rate": 7.708351696782724e-06,
      "loss": 4.068466796875,
      "step": 21800
    },
    {
      "epoch": 9.590428431265671,
      "grad_norm": 24.918800354003906,
      "learning_rate": 7.686315557514325e-06,
      "loss": 4.066632690429688,
      "step": 22000
    },
    {
      "epoch": 9.590428431265671,
      "eval_loss": 0.9627618193626404,
      "eval_runtime": 353.7365,
      "eval_samples_per_second": 34.577,
      "eval_steps_per_second": 8.645,
      "step": 22000
    },
    {
      "epoch": 9.67764090264908,
      "grad_norm": 21.744169235229492,
      "learning_rate": 7.664279418245924e-06,
      "loss": 4.0513720703125,
      "step": 22200
    },
    {
      "epoch": 9.764853374032487,
      "grad_norm": 20.75674057006836,
      "learning_rate": 7.642243278977523e-06,
      "loss": 4.0559814453125,
      "step": 22400
    },
    {
      "epoch": 9.852065845415895,
      "grad_norm": 38.736656188964844,
      "learning_rate": 7.620207139709123e-06,
      "loss": 4.05899658203125,
      "step": 22600
    },
    {
      "epoch": 9.939278316799303,
      "grad_norm": 24.679214477539062,
      "learning_rate": 7.598171000440724e-06,
      "loss": 4.065103759765625,
      "step": 22800
    },
    {
      "epoch": 10.026163741415022,
      "grad_norm": 20.022254943847656,
      "learning_rate": 7.576134861172323e-06,
      "loss": 4.0568896484375,
      "step": 23000
    },
    {
      "epoch": 10.026163741415022,
      "eval_loss": 0.9595907926559448,
      "eval_runtime": 323.4808,
      "eval_samples_per_second": 37.811,
      "eval_steps_per_second": 9.453,
      "step": 23000
    },
    {
      "epoch": 10.11337621279843,
      "grad_norm": 23.9129695892334,
      "learning_rate": 7.554098721903923e-06,
      "loss": 4.054570922851562,
      "step": 23200
    },
    {
      "epoch": 10.200588684181838,
      "grad_norm": 27.955625534057617,
      "learning_rate": 7.532062582635523e-06,
      "loss": 4.035859069824219,
      "step": 23400
    },
    {
      "epoch": 10.287801155565246,
      "grad_norm": 31.21686363220215,
      "learning_rate": 7.510026443367122e-06,
      "loss": 4.056401977539062,
      "step": 23600
    },
    {
      "epoch": 10.375013626948654,
      "grad_norm": 37.817901611328125,
      "learning_rate": 7.487990304098722e-06,
      "loss": 4.037947387695312,
      "step": 23800
    },
    {
      "epoch": 10.462226098332062,
      "grad_norm": 19.933561325073242,
      "learning_rate": 7.465954164830322e-06,
      "loss": 4.054637451171875,
      "step": 24000
    },
    {
      "epoch": 10.462226098332062,
      "eval_loss": 0.961605429649353,
      "eval_runtime": 338.6319,
      "eval_samples_per_second": 36.119,
      "eval_steps_per_second": 9.03,
      "step": 24000
    },
    {
      "epoch": 10.54943856971547,
      "grad_norm": 18.578306198120117,
      "learning_rate": 7.443918025561922e-06,
      "loss": 4.040963745117187,
      "step": 24200
    },
    {
      "epoch": 10.636651041098878,
      "grad_norm": 20.99644660949707,
      "learning_rate": 7.421881886293521e-06,
      "loss": 4.051210021972656,
      "step": 24400
    },
    {
      "epoch": 10.723863512482286,
      "grad_norm": 31.51519203186035,
      "learning_rate": 7.399845747025122e-06,
      "loss": 4.038525390625,
      "step": 24600
    },
    {
      "epoch": 10.811075983865694,
      "grad_norm": 24.11183738708496,
      "learning_rate": 7.377809607756722e-06,
      "loss": 4.033809204101562,
      "step": 24800
    },
    {
      "epoch": 10.898288455249102,
      "grad_norm": 20.393661499023438,
      "learning_rate": 7.355773468488322e-06,
      "loss": 4.0559149169921875,
      "step": 25000
    },
    {
      "epoch": 10.898288455249102,
      "eval_loss": 0.9618766903877258,
      "eval_runtime": 334.0788,
      "eval_samples_per_second": 36.611,
      "eval_steps_per_second": 9.154,
      "step": 25000
    },
    {
      "epoch": 10.985500926632508,
      "grad_norm": 29.87595558166504,
      "learning_rate": 7.333737329219921e-06,
      "loss": 4.033812255859375,
      "step": 25200
    },
    {
      "epoch": 11.072386351248229,
      "grad_norm": 21.925132751464844,
      "learning_rate": 7.311701189951521e-06,
      "loss": 4.00488037109375,
      "step": 25400
    },
    {
      "epoch": 11.159598822631636,
      "grad_norm": 26.820737838745117,
      "learning_rate": 7.289665050683121e-06,
      "loss": 4.017337646484375,
      "step": 25600
    },
    {
      "epoch": 11.246811294015044,
      "grad_norm": 20.507617950439453,
      "learning_rate": 7.26762891141472e-06,
      "loss": 4.037488403320313,
      "step": 25800
    },
    {
      "epoch": 11.334023765398452,
      "grad_norm": 29.21367835998535,
      "learning_rate": 7.24559277214632e-06,
      "loss": 4.0288302612304685,
      "step": 26000
    },
    {
      "epoch": 11.334023765398452,
      "eval_loss": 0.9569966793060303,
      "eval_runtime": 329.0449,
      "eval_samples_per_second": 37.171,
      "eval_steps_per_second": 9.294,
      "step": 26000
    },
    {
      "epoch": 11.42123623678186,
      "grad_norm": 20.909095764160156,
      "learning_rate": 7.223556632877921e-06,
      "loss": 4.025175170898438,
      "step": 26200
    },
    {
      "epoch": 11.508448708165268,
      "grad_norm": 23.773263931274414,
      "learning_rate": 7.201520493609521e-06,
      "loss": 4.032984619140625,
      "step": 26400
    },
    {
      "epoch": 11.595661179548676,
      "grad_norm": 31.082735061645508,
      "learning_rate": 7.17948435434112e-06,
      "loss": 4.03861572265625,
      "step": 26600
    },
    {
      "epoch": 11.682873650932084,
      "grad_norm": 22.946895599365234,
      "learning_rate": 7.15744821507272e-06,
      "loss": 4.023303833007812,
      "step": 26800
    },
    {
      "epoch": 11.77008612231549,
      "grad_norm": 24.374385833740234,
      "learning_rate": 7.13541207580432e-06,
      "loss": 4.0240914916992185,
      "step": 27000
    },
    {
      "epoch": 11.77008612231549,
      "eval_loss": 0.9545909762382507,
      "eval_runtime": 319.1017,
      "eval_samples_per_second": 38.329,
      "eval_steps_per_second": 9.583,
      "step": 27000
    },
    {
      "epoch": 11.857298593698898,
      "grad_norm": 28.567989349365234,
      "learning_rate": 7.113375936535919e-06,
      "loss": 4.023042297363281,
      "step": 27200
    },
    {
      "epoch": 11.944511065082306,
      "grad_norm": 27.42933464050293,
      "learning_rate": 7.091339797267519e-06,
      "loss": 4.001913146972656,
      "step": 27400
    },
    {
      "epoch": 12.031396489698027,
      "grad_norm": 28.52174949645996,
      "learning_rate": 7.069303657999119e-06,
      "loss": 4.006860046386719,
      "step": 27600
    },
    {
      "epoch": 12.118608961081435,
      "grad_norm": 23.863718032836914,
      "learning_rate": 7.04726751873072e-06,
      "loss": 4.015152587890625,
      "step": 27800
    },
    {
      "epoch": 12.205821432464843,
      "grad_norm": 26.30300521850586,
      "learning_rate": 7.025231379462319e-06,
      "loss": 3.998001708984375,
      "step": 28000
    },
    {
      "epoch": 12.205821432464843,
      "eval_loss": 0.9601166844367981,
      "eval_runtime": 325.8118,
      "eval_samples_per_second": 37.54,
      "eval_steps_per_second": 9.386,
      "step": 28000
    },
    {
      "epoch": 12.29303390384825,
      "grad_norm": 26.98280906677246,
      "learning_rate": 7.003195240193919e-06,
      "loss": 4.022799682617188,
      "step": 28200
    },
    {
      "epoch": 12.380246375231659,
      "grad_norm": 30.50926971435547,
      "learning_rate": 6.981159100925519e-06,
      "loss": 4.020477294921875,
      "step": 28400
    },
    {
      "epoch": 12.467458846615067,
      "grad_norm": 18.858837127685547,
      "learning_rate": 6.959122961657119e-06,
      "loss": 4.030046997070312,
      "step": 28600
    },
    {
      "epoch": 12.554671317998475,
      "grad_norm": 33.04899597167969,
      "learning_rate": 6.937086822388718e-06,
      "loss": 4.016376037597656,
      "step": 28800
    },
    {
      "epoch": 12.641883789381883,
      "grad_norm": 17.96835708618164,
      "learning_rate": 6.915050683120318e-06,
      "loss": 4.029109191894531,
      "step": 29000
    },
    {
      "epoch": 12.641883789381883,
      "eval_loss": 0.9563982486724854,
      "eval_runtime": 326.5266,
      "eval_samples_per_second": 37.458,
      "eval_steps_per_second": 9.365,
      "step": 29000
    },
    {
      "epoch": 12.729096260765289,
      "grad_norm": 21.767717361450195,
      "learning_rate": 6.8930145438519185e-06,
      "loss": 4.005764770507812,
      "step": 29200
    },
    {
      "epoch": 12.816308732148697,
      "grad_norm": 27.48580551147461,
      "learning_rate": 6.870978404583517e-06,
      "loss": 4.0095166015625,
      "step": 29400
    },
    {
      "epoch": 12.903521203532105,
      "grad_norm": 24.953073501586914,
      "learning_rate": 6.8489422653151175e-06,
      "loss": 4.012142028808594,
      "step": 29600
    },
    {
      "epoch": 12.990733674915512,
      "grad_norm": 22.37385368347168,
      "learning_rate": 6.8269061260467175e-06,
      "loss": 4.005128173828125,
      "step": 29800
    },
    {
      "epoch": 13.077619099531233,
      "grad_norm": 27.91280174255371,
      "learning_rate": 6.8048699867783174e-06,
      "loss": 3.9836300659179686,
      "step": 30000
    },
    {
      "epoch": 13.077619099531233,
      "eval_loss": 0.9551950097084045,
      "eval_runtime": 323.8874,
      "eval_samples_per_second": 37.763,
      "eval_steps_per_second": 9.442,
      "step": 30000
    },
    {
      "epoch": 13.164831570914641,
      "grad_norm": 21.42306900024414,
      "learning_rate": 6.7828338475099165e-06,
      "loss": 3.9929656982421875,
      "step": 30200
    },
    {
      "epoch": 13.25204404229805,
      "grad_norm": 22.932331085205078,
      "learning_rate": 6.7607977082415165e-06,
      "loss": 3.995915832519531,
      "step": 30400
    },
    {
      "epoch": 13.339256513681457,
      "grad_norm": 19.06012725830078,
      "learning_rate": 6.7387615689731164e-06,
      "loss": 4.001346435546875,
      "step": 30600
    },
    {
      "epoch": 13.426468985064865,
      "grad_norm": 28.957733154296875,
      "learning_rate": 6.7167254297047156e-06,
      "loss": 4.010101318359375,
      "step": 30800
    },
    {
      "epoch": 13.513681456448271,
      "grad_norm": 33.10238265991211,
      "learning_rate": 6.6946892904363155e-06,
      "loss": 4.0022412109375,
      "step": 31000
    },
    {
      "epoch": 13.513681456448271,
      "eval_loss": 0.955825686454773,
      "eval_runtime": 324.5053,
      "eval_samples_per_second": 37.691,
      "eval_steps_per_second": 9.424,
      "step": 31000
    },
    {
      "epoch": 13.60089392783168,
      "grad_norm": 30.01372718811035,
      "learning_rate": 6.672653151167916e-06,
      "loss": 4.007727355957031,
      "step": 31200
    },
    {
      "epoch": 13.688106399215087,
      "grad_norm": 21.847442626953125,
      "learning_rate": 6.650617011899516e-06,
      "loss": 3.993216247558594,
      "step": 31400
    },
    {
      "epoch": 13.775318870598495,
      "grad_norm": 31.029464721679688,
      "learning_rate": 6.628580872631115e-06,
      "loss": 3.9984600830078123,
      "step": 31600
    },
    {
      "epoch": 13.862531341981903,
      "grad_norm": 39.60411834716797,
      "learning_rate": 6.606544733362715e-06,
      "loss": 4.00242919921875,
      "step": 31800
    },
    {
      "epoch": 13.94974381336531,
      "grad_norm": 23.86924171447754,
      "learning_rate": 6.584508594094315e-06,
      "loss": 3.9853219604492187,
      "step": 32000
    },
    {
      "epoch": 13.94974381336531,
      "eval_loss": 0.9622883200645447,
      "eval_runtime": 341.701,
      "eval_samples_per_second": 35.794,
      "eval_steps_per_second": 8.949,
      "step": 32000
    },
    {
      "epoch": 14.036629237981032,
      "grad_norm": 28.95281219482422,
      "learning_rate": 6.562472454825915e-06,
      "loss": 3.991866149902344,
      "step": 32200
    },
    {
      "epoch": 14.12384170936444,
      "grad_norm": 26.760395050048828,
      "learning_rate": 6.540436315557514e-06,
      "loss": 4.00505859375,
      "step": 32400
    },
    {
      "epoch": 14.211054180747848,
      "grad_norm": 26.774749755859375,
      "learning_rate": 6.518400176289114e-06,
      "loss": 3.9822613525390627,
      "step": 32600
    },
    {
      "epoch": 14.298266652131256,
      "grad_norm": 23.230669021606445,
      "learning_rate": 6.496364037020715e-06,
      "loss": 4.0000741577148435,
      "step": 32800
    },
    {
      "epoch": 14.385479123514662,
      "grad_norm": 24.043460845947266,
      "learning_rate": 6.474327897752314e-06,
      "loss": 3.9965869140625,
      "step": 33000
    },
    {
      "epoch": 14.385479123514662,
      "eval_loss": 0.9531711339950562,
      "eval_runtime": 383.9679,
      "eval_samples_per_second": 31.854,
      "eval_steps_per_second": 7.964,
      "step": 33000
    },
    {
      "epoch": 14.47269159489807,
      "grad_norm": 20.8858699798584,
      "learning_rate": 6.452291758483914e-06,
      "loss": 3.986736145019531,
      "step": 33200
    },
    {
      "epoch": 14.559904066281478,
      "grad_norm": 23.316043853759766,
      "learning_rate": 6.430255619215514e-06,
      "loss": 3.9917340087890625,
      "step": 33400
    },
    {
      "epoch": 14.647116537664886,
      "grad_norm": 29.799169540405273,
      "learning_rate": 6.408219479947114e-06,
      "loss": 3.9852001953125,
      "step": 33600
    },
    {
      "epoch": 14.734329009048293,
      "grad_norm": 27.449323654174805,
      "learning_rate": 6.386183340678713e-06,
      "loss": 3.9847030639648438,
      "step": 33800
    },
    {
      "epoch": 14.821541480431701,
      "grad_norm": 33.31381607055664,
      "learning_rate": 6.364147201410313e-06,
      "loss": 3.9803167724609376,
      "step": 34000
    },
    {
      "epoch": 14.821541480431701,
      "eval_loss": 0.9532381892204285,
      "eval_runtime": 382.6451,
      "eval_samples_per_second": 31.964,
      "eval_steps_per_second": 7.992,
      "step": 34000
    },
    {
      "epoch": 14.90875395181511,
      "grad_norm": 28.379302978515625,
      "learning_rate": 6.342111062141914e-06,
      "loss": 3.97963134765625,
      "step": 34200
    },
    {
      "epoch": 14.995966423198517,
      "grad_norm": 23.981943130493164,
      "learning_rate": 6.320074922873512e-06,
      "loss": 3.9923846435546877,
      "step": 34400
    },
    {
      "epoch": 15.082851847814238,
      "grad_norm": 21.363521575927734,
      "learning_rate": 6.298038783605113e-06,
      "loss": 3.9748077392578125,
      "step": 34600
    },
    {
      "epoch": 15.170064319197646,
      "grad_norm": 27.216100692749023,
      "learning_rate": 6.276002644336713e-06,
      "loss": 3.9767276000976564,
      "step": 34800
    },
    {
      "epoch": 15.257276790581052,
      "grad_norm": 26.439565658569336,
      "learning_rate": 6.253966505068313e-06,
      "loss": 3.9838623046875,
      "step": 35000
    },
    {
      "epoch": 15.257276790581052,
      "eval_loss": 0.951134204864502,
      "eval_runtime": 382.1653,
      "eval_samples_per_second": 32.004,
      "eval_steps_per_second": 8.002,
      "step": 35000
    },
    {
      "epoch": 15.34448926196446,
      "grad_norm": 27.999643325805664,
      "learning_rate": 6.231930365799912e-06,
      "loss": 3.978865966796875,
      "step": 35200
    },
    {
      "epoch": 15.431701733347868,
      "grad_norm": 23.71980857849121,
      "learning_rate": 6.209894226531512e-06,
      "loss": 3.9701495361328125,
      "step": 35400
    },
    {
      "epoch": 15.518914204731276,
      "grad_norm": 31.778179168701172,
      "learning_rate": 6.187858087263112e-06,
      "loss": 3.971527404785156,
      "step": 35600
    },
    {
      "epoch": 15.606126676114684,
      "grad_norm": 27.436248779296875,
      "learning_rate": 6.165821947994713e-06,
      "loss": 3.9800680541992186,
      "step": 35800
    },
    {
      "epoch": 15.693339147498092,
      "grad_norm": 25.100250244140625,
      "learning_rate": 6.143785808726311e-06,
      "loss": 3.975406494140625,
      "step": 36000
    },
    {
      "epoch": 15.693339147498092,
      "eval_loss": 0.9510881900787354,
      "eval_runtime": 387.6496,
      "eval_samples_per_second": 31.552,
      "eval_steps_per_second": 7.889,
      "step": 36000
    },
    {
      "epoch": 15.7805516188815,
      "grad_norm": 24.325748443603516,
      "learning_rate": 6.121749669457912e-06,
      "loss": 3.9792324829101564,
      "step": 36200
    },
    {
      "epoch": 15.867764090264908,
      "grad_norm": 25.83987808227539,
      "learning_rate": 6.099713530189512e-06,
      "loss": 3.96053466796875,
      "step": 36400
    },
    {
      "epoch": 15.954976561648316,
      "grad_norm": 43.17454147338867,
      "learning_rate": 6.077677390921111e-06,
      "loss": 3.981627197265625,
      "step": 36600
    },
    {
      "epoch": 16.041861986264035,
      "grad_norm": 28.25941276550293,
      "learning_rate": 6.055641251652711e-06,
      "loss": 3.948244323730469,
      "step": 36800
    },
    {
      "epoch": 16.129074457647445,
      "grad_norm": 32.71881866455078,
      "learning_rate": 6.033605112384311e-06,
      "loss": 3.977437744140625,
      "step": 37000
    },
    {
      "epoch": 16.129074457647445,
      "eval_loss": 0.9528171420097351,
      "eval_runtime": 385.8089,
      "eval_samples_per_second": 31.702,
      "eval_steps_per_second": 7.926,
      "step": 37000
    },
    {
      "epoch": 16.21628692903085,
      "grad_norm": 24.57404136657715,
      "learning_rate": 6.011568973115911e-06,
      "loss": 3.9684487915039064,
      "step": 37200
    },
    {
      "epoch": 16.30349940041426,
      "grad_norm": 21.916828155517578,
      "learning_rate": 5.98953283384751e-06,
      "loss": 3.9711444091796877,
      "step": 37400
    },
    {
      "epoch": 16.390711871797667,
      "grad_norm": 26.631895065307617,
      "learning_rate": 5.96749669457911e-06,
      "loss": 3.9632708740234377,
      "step": 37600
    },
    {
      "epoch": 16.477924343181076,
      "grad_norm": 33.824806213378906,
      "learning_rate": 5.9454605553107106e-06,
      "loss": 3.9652227783203124,
      "step": 37800
    },
    {
      "epoch": 16.565136814564482,
      "grad_norm": 25.653230667114258,
      "learning_rate": 5.92342441604231e-06,
      "loss": 3.9611251831054686,
      "step": 38000
    },
    {
      "epoch": 16.565136814564482,
      "eval_loss": 0.9488371014595032,
      "eval_runtime": 402.1828,
      "eval_samples_per_second": 30.412,
      "eval_steps_per_second": 7.604,
      "step": 38000
    },
    {
      "epoch": 16.652349285947892,
      "grad_norm": 23.87160301208496,
      "learning_rate": 5.90138827677391e-06,
      "loss": 3.968856201171875,
      "step": 38200
    },
    {
      "epoch": 16.7395617573313,
      "grad_norm": 29.487102508544922,
      "learning_rate": 5.87935213750551e-06,
      "loss": 3.9722393798828124,
      "step": 38400
    },
    {
      "epoch": 16.826774228714704,
      "grad_norm": 30.376422882080078,
      "learning_rate": 5.8573159982371095e-06,
      "loss": 3.9577349853515624,
      "step": 38600
    },
    {
      "epoch": 16.913986700098114,
      "grad_norm": 25.18212127685547,
      "learning_rate": 5.835279858968709e-06,
      "loss": 3.9670977783203125,
      "step": 38800
    },
    {
      "epoch": 17.000872124713833,
      "grad_norm": 22.56792449951172,
      "learning_rate": 5.813243719700309e-06,
      "loss": 3.9620367431640626,
      "step": 39000
    },
    {
      "epoch": 17.000872124713833,
      "eval_loss": 0.9485888481140137,
      "eval_runtime": 354.5395,
      "eval_samples_per_second": 34.498,
      "eval_steps_per_second": 8.625,
      "step": 39000
    },
    {
      "epoch": 17.088084596097243,
      "grad_norm": 26.586441040039062,
      "learning_rate": 5.791207580431909e-06,
      "loss": 3.9551470947265623,
      "step": 39200
    },
    {
      "epoch": 17.17529706748065,
      "grad_norm": 21.796438217163086,
      "learning_rate": 5.769171441163509e-06,
      "loss": 3.942110290527344,
      "step": 39400
    },
    {
      "epoch": 17.26250953886406,
      "grad_norm": 22.50459098815918,
      "learning_rate": 5.7471353018951085e-06,
      "loss": 3.9705413818359374,
      "step": 39600
    },
    {
      "epoch": 17.349722010247465,
      "grad_norm": 22.940710067749023,
      "learning_rate": 5.725099162626708e-06,
      "loss": 3.96304931640625,
      "step": 39800
    },
    {
      "epoch": 17.436934481630875,
      "grad_norm": 23.740083694458008,
      "learning_rate": 5.703063023358308e-06,
      "loss": 3.9483935546875,
      "step": 40000
    },
    {
      "epoch": 17.436934481630875,
      "eval_loss": 0.9485511779785156,
      "eval_runtime": 389.0835,
      "eval_samples_per_second": 31.435,
      "eval_steps_per_second": 7.859,
      "step": 40000
    },
    {
      "epoch": 17.52414695301428,
      "grad_norm": 20.306228637695312,
      "learning_rate": 5.6810268840899075e-06,
      "loss": 3.9628460693359373,
      "step": 40200
    },
    {
      "epoch": 17.61135942439769,
      "grad_norm": 21.633869171142578,
      "learning_rate": 5.6589907448215074e-06,
      "loss": 3.955937805175781,
      "step": 40400
    },
    {
      "epoch": 17.698571895781097,
      "grad_norm": 31.192277908325195,
      "learning_rate": 5.636954605553107e-06,
      "loss": 3.965909423828125,
      "step": 40600
    },
    {
      "epoch": 17.785784367164503,
      "grad_norm": 22.128379821777344,
      "learning_rate": 5.614918466284708e-06,
      "loss": 3.967447509765625,
      "step": 40800
    },
    {
      "epoch": 17.872996838547913,
      "grad_norm": 47.48439407348633,
      "learning_rate": 5.5928823270163065e-06,
      "loss": 3.954619140625,
      "step": 41000
    },
    {
      "epoch": 17.872996838547913,
      "eval_loss": 0.9497806429862976,
      "eval_runtime": 369.1518,
      "eval_samples_per_second": 33.133,
      "eval_steps_per_second": 8.284,
      "step": 41000
    },
    {
      "epoch": 17.96020930993132,
      "grad_norm": 37.53486251831055,
      "learning_rate": 5.570846187747907e-06,
      "loss": 3.959051513671875,
      "step": 41200
    },
    {
      "epoch": 18.04709473454704,
      "grad_norm": 33.98725891113281,
      "learning_rate": 5.548810048479507e-06,
      "loss": 3.937330322265625,
      "step": 41400
    },
    {
      "epoch": 18.134307205930448,
      "grad_norm": 25.65367317199707,
      "learning_rate": 5.526773909211106e-06,
      "loss": 3.947315673828125,
      "step": 41600
    },
    {
      "epoch": 18.221519677313857,
      "grad_norm": 27.725730895996094,
      "learning_rate": 5.504737769942706e-06,
      "loss": 3.957863464355469,
      "step": 41800
    },
    {
      "epoch": 18.308732148697263,
      "grad_norm": 30.85365104675293,
      "learning_rate": 5.482701630674306e-06,
      "loss": 3.950794982910156,
      "step": 42000
    },
    {
      "epoch": 18.308732148697263,
      "eval_loss": 0.9448174834251404,
      "eval_runtime": 358.5243,
      "eval_samples_per_second": 34.115,
      "eval_steps_per_second": 8.529,
      "step": 42000
    }
  ],
  "logging_steps": 200,
  "max_steps": 91760,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 40,
  "save_steps": 1000,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 2.783882407575552e+16,
  "train_batch_size": 6,
  "trial_name": null,
  "trial_params": null
}