{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.0,
  "eval_steps": 250,
  "global_step": 2675,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.05009345794392523,
      "grad_norm": 0.6554287672042847,
      "learning_rate": 5e-07,
      "loss": 0.4929,
      "step": 134
    },
    {
      "epoch": 0.09345794392523364,
      "eval_sft_sg_values_eval_loss": 0.4027545750141144,
      "eval_sft_sg_values_eval_runtime": 28.7907,
      "eval_sft_sg_values_eval_samples_per_second": 17.367,
      "eval_sft_sg_values_eval_steps_per_second": 8.683,
      "step": 250
    },
    {
      "epoch": 0.10018691588785046,
      "grad_norm": 0.877953827381134,
      "learning_rate": 1e-06,
      "loss": 0.4862,
      "step": 268
    },
    {
      "epoch": 0.1502803738317757,
      "grad_norm": 1.139445185661316,
      "learning_rate": 9.443290402991276e-07,
      "loss": 0.4112,
      "step": 402
    },
    {
      "epoch": 0.18691588785046728,
      "eval_sft_sg_values_eval_loss": 0.35132163763046265,
      "eval_sft_sg_values_eval_runtime": 28.7753,
      "eval_sft_sg_values_eval_samples_per_second": 17.376,
      "eval_sft_sg_values_eval_steps_per_second": 8.688,
      "step": 500
    },
    {
      "epoch": 0.20037383177570092,
      "grad_norm": 1.0352528095245361,
      "learning_rate": 8.886580805982551e-07,
      "loss": 0.3246,
      "step": 536
    },
    {
      "epoch": 0.2504672897196262,
      "grad_norm": 1.3118387460708618,
      "learning_rate": 8.329871208973826e-07,
      "loss": 0.2797,
      "step": 670
    },
    {
      "epoch": 0.2803738317757009,
      "eval_sft_sg_values_eval_loss": 0.2957317531108856,
      "eval_sft_sg_values_eval_runtime": 28.7598,
      "eval_sft_sg_values_eval_samples_per_second": 17.385,
      "eval_sft_sg_values_eval_steps_per_second": 8.693,
      "step": 750
    },
    {
      "epoch": 0.3005607476635514,
      "grad_norm": 1.1230024099349976,
      "learning_rate": 7.773161611965101e-07,
      "loss": 0.2643,
      "step": 804
    },
    {
      "epoch": 0.3506542056074766,
      "grad_norm": 1.5997352600097656,
      "learning_rate": 7.216452014956377e-07,
      "loss": 0.2602,
      "step": 938
    },
    {
      "epoch": 0.37383177570093457,
      "eval_sft_sg_values_eval_loss": 0.26124340295791626,
      "eval_sft_sg_values_eval_runtime": 28.7839,
      "eval_sft_sg_values_eval_samples_per_second": 17.371,
      "eval_sft_sg_values_eval_steps_per_second": 8.685,
      "step": 1000
    },
    {
      "epoch": 0.40074766355140184,
      "grad_norm": 2.0844686031341553,
      "learning_rate": 6.659742417947652e-07,
      "loss": 0.2542,
      "step": 1072
    },
    {
      "epoch": 0.4508411214953271,
      "grad_norm": 1.0873792171478271,
      "learning_rate": 6.103032820938928e-07,
      "loss": 0.2503,
      "step": 1206
    },
    {
      "epoch": 0.4672897196261682,
      "eval_sft_sg_values_eval_loss": 0.23845288157463074,
      "eval_sft_sg_values_eval_runtime": 28.7889,
      "eval_sft_sg_values_eval_samples_per_second": 17.368,
      "eval_sft_sg_values_eval_steps_per_second": 8.684,
      "step": 1250
    },
    {
      "epoch": 0.5009345794392523,
      "grad_norm": 1.5952402353286743,
      "learning_rate": 5.546323223930204e-07,
      "loss": 0.2441,
      "step": 1340
    },
    {
      "epoch": 0.5510280373831775,
      "grad_norm": 1.4540523290634155,
      "learning_rate": 4.989613626921479e-07,
      "loss": 0.2416,
      "step": 1474
    },
    {
      "epoch": 0.5607476635514018,
      "eval_sft_sg_values_eval_loss": 0.22509750723838806,
      "eval_sft_sg_values_eval_runtime": 28.785,
      "eval_sft_sg_values_eval_samples_per_second": 17.37,
      "eval_sft_sg_values_eval_steps_per_second": 8.685,
      "step": 1500
    },
    {
      "epoch": 0.6011214953271028,
      "grad_norm": 1.5883971452713013,
      "learning_rate": 4.4329040299127543e-07,
      "loss": 0.2381,
      "step": 1608
    },
    {
      "epoch": 0.6512149532710281,
      "grad_norm": 1.1363577842712402,
      "learning_rate": 3.87619443290403e-07,
      "loss": 0.238,
      "step": 1742
    },
    {
      "epoch": 0.6542056074766355,
      "eval_sft_sg_values_eval_loss": 0.2166362851858139,
      "eval_sft_sg_values_eval_runtime": 28.8027,
      "eval_sft_sg_values_eval_samples_per_second": 17.359,
      "eval_sft_sg_values_eval_steps_per_second": 8.68,
      "step": 1750
    },
    {
      "epoch": 0.7013084112149532,
      "grad_norm": 1.0956478118896484,
      "learning_rate": 3.3194848358953054e-07,
      "loss": 0.2332,
      "step": 1876
    },
    {
      "epoch": 0.7476635514018691,
      "eval_sft_sg_values_eval_loss": 0.21183913946151733,
      "eval_sft_sg_values_eval_runtime": 28.802,
      "eval_sft_sg_values_eval_samples_per_second": 17.36,
      "eval_sft_sg_values_eval_steps_per_second": 8.68,
      "step": 2000
    },
    {
      "epoch": 0.7514018691588785,
      "grad_norm": 1.294016718864441,
      "learning_rate": 2.7627752388865805e-07,
      "loss": 0.2282,
      "step": 2010
    },
    {
      "epoch": 0.8014953271028037,
      "grad_norm": 1.3785642385482788,
      "learning_rate": 2.206065641877856e-07,
      "loss": 0.2335,
      "step": 2144
    },
    {
      "epoch": 0.8411214953271028,
      "eval_sft_sg_values_eval_loss": 0.20833030343055725,
      "eval_sft_sg_values_eval_runtime": 28.8236,
      "eval_sft_sg_values_eval_samples_per_second": 17.347,
      "eval_sft_sg_values_eval_steps_per_second": 8.673,
      "step": 2250
    },
    {
      "epoch": 0.851588785046729,
      "grad_norm": 1.1724706888198853,
      "learning_rate": 1.6493560448691317e-07,
      "loss": 0.2307,
      "step": 2278
    },
    {
      "epoch": 0.9016822429906542,
      "grad_norm": 2.404283285140991,
      "learning_rate": 1.0926464478604071e-07,
      "loss": 0.2359,
      "step": 2412
    },
    {
      "epoch": 0.9345794392523364,
      "eval_sft_sg_values_eval_loss": 0.20680196583271027,
      "eval_sft_sg_values_eval_runtime": 28.7876,
      "eval_sft_sg_values_eval_samples_per_second": 17.369,
      "eval_sft_sg_values_eval_steps_per_second": 8.684,
      "step": 2500
    },
    {
      "epoch": 0.9517757009345794,
      "grad_norm": 1.2437078952789307,
      "learning_rate": 5.3593685085168255e-08,
      "loss": 0.2313,
      "step": 2546
    },
    {
      "epoch": 1.0,
      "step": 2675,
      "total_flos": 3.769390263140352e+17,
      "train_loss": 0.280462795507128,
      "train_runtime": 7587.9769,
      "train_samples_per_second": 2.82,
      "train_steps_per_second": 0.353
    }
  ],
  "logging_steps": 134,
  "max_steps": 2675,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 3.769390263140352e+17,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}