{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.0,
  "eval_steps": 500,
  "global_step": 8592,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.00011638733705772812,
      "grad_norm": 8.116299629211426,
      "learning_rate": 0.0,
      "loss": 2.7266,
      "step": 1
    },
    {
      "epoch": 0.05819366852886406,
      "grad_norm": 57.76313781738281,
      "learning_rate": 1.732045163543243e-06,
      "loss": 3.7373,
      "step": 500
    },
    {
      "epoch": 0.05819366852886406,
      "eval_cosine_accuracy": 0.9404736757278442,
      "eval_loss": 2.536726713180542,
      "eval_runtime": 26.2292,
      "eval_samples_per_second": 360.59,
      "eval_steps_per_second": 1.411,
      "step": 500
    },
    {
      "epoch": 0.11638733705772812,
      "grad_norm": 5.569820880889893,
      "learning_rate": 3.4780584332440925e-06,
      "loss": 2.8655,
      "step": 1000
    },
    {
      "epoch": 0.11638733705772812,
      "eval_cosine_accuracy": 0.9364559054374695,
      "eval_loss": 2.6046745777130127,
      "eval_runtime": 26.7465,
      "eval_samples_per_second": 353.617,
      "eval_steps_per_second": 1.383,
      "step": 1000
    },
    {
      "epoch": 0.17458100558659218,
      "grad_norm": 5.892590045928955,
      "learning_rate": 5.224071702944943e-06,
      "loss": 2.3859,
      "step": 1500
    },
    {
      "epoch": 0.17458100558659218,
      "eval_cosine_accuracy": 0.9412137866020203,
      "eval_loss": 2.540637969970703,
      "eval_runtime": 26.8122,
      "eval_samples_per_second": 352.75,
      "eval_steps_per_second": 1.38,
      "step": 1500
    },
    {
      "epoch": 0.23277467411545624,
      "grad_norm": 41.33493423461914,
      "learning_rate": 6.970084972645793e-06,
      "loss": 2.1884,
      "step": 2000
    },
    {
      "epoch": 0.23277467411545624,
      "eval_cosine_accuracy": 0.9451258182525635,
      "eval_loss": 2.5318424701690674,
      "eval_runtime": 26.0188,
      "eval_samples_per_second": 363.506,
      "eval_steps_per_second": 1.422,
      "step": 2000
    },
    {
      "epoch": 0.2909683426443203,
      "grad_norm": 15.76939582824707,
      "learning_rate": 8.71260621580724e-06,
      "loss": 1.9576,
      "step": 2500
    },
    {
      "epoch": 0.2909683426443203,
      "eval_cosine_accuracy": 0.9467117786407471,
      "eval_loss": 2.474266767501831,
      "eval_runtime": 26.2939,
      "eval_samples_per_second": 359.703,
      "eval_steps_per_second": 1.407,
      "step": 2500
    },
    {
      "epoch": 0.34916201117318435,
      "grad_norm": 7.4116950035095215,
      "learning_rate": 1.045861948550809e-05,
      "loss": 1.8211,
      "step": 3000
    },
    {
      "epoch": 0.34916201117318435,
      "eval_cosine_accuracy": 0.9435398578643799,
      "eval_loss": 2.547072649002075,
      "eval_runtime": 26.1008,
      "eval_samples_per_second": 362.365,
      "eval_steps_per_second": 1.418,
      "step": 3000
    },
    {
      "epoch": 0.4073556797020484,
      "grad_norm": 25.99747657775879,
      "learning_rate": 1.2204632755208939e-05,
      "loss": 1.6603,
      "step": 3500
    },
    {
      "epoch": 0.4073556797020484,
      "eval_cosine_accuracy": 0.9489321112632751,
      "eval_loss": 2.472174644470215,
      "eval_runtime": 27.1444,
      "eval_samples_per_second": 348.433,
      "eval_steps_per_second": 1.363,
      "step": 3500
    },
    {
      "epoch": 0.4655493482309125,
      "grad_norm": 5.763104438781738,
      "learning_rate": 1.395064602490979e-05,
      "loss": 1.596,
      "step": 4000
    },
    {
      "epoch": 0.4655493482309125,
      "eval_cosine_accuracy": 0.9438570737838745,
      "eval_loss": 2.5426251888275146,
      "eval_runtime": 28.2718,
      "eval_samples_per_second": 334.538,
      "eval_steps_per_second": 1.309,
      "step": 4000
    },
    {
      "epoch": 0.5237430167597765,
      "grad_norm": 4.893315315246582,
      "learning_rate": 1.5693167268071237e-05,
      "loss": 1.5379,
      "step": 4500
    },
    {
      "epoch": 0.5237430167597765,
      "eval_cosine_accuracy": 0.9473461508750916,
      "eval_loss": 2.4768149852752686,
      "eval_runtime": 28.2378,
      "eval_samples_per_second": 334.941,
      "eval_steps_per_second": 1.31,
      "step": 4500
    },
    {
      "epoch": 0.5819366852886406,
      "grad_norm": 43.368614196777344,
      "learning_rate": 1.7439180537772086e-05,
      "loss": 1.5397,
      "step": 5000
    },
    {
      "epoch": 0.5819366852886406,
      "eval_cosine_accuracy": 0.9487206339836121,
      "eval_loss": 2.4771170616149902,
      "eval_runtime": 28.4116,
      "eval_samples_per_second": 332.892,
      "eval_steps_per_second": 1.302,
      "step": 5000
    },
    {
      "epoch": 0.6401303538175046,
      "grad_norm": 0.21546457707881927,
      "learning_rate": 1.9185193807472936e-05,
      "loss": 1.381,
      "step": 5500
    },
    {
      "epoch": 0.6401303538175046,
      "eval_cosine_accuracy": 0.9412137866020203,
      "eval_loss": 2.6126925945281982,
      "eval_runtime": 28.5708,
      "eval_samples_per_second": 331.037,
      "eval_steps_per_second": 1.295,
      "step": 5500
    },
    {
      "epoch": 0.6983240223463687,
      "grad_norm": 27.70214080810547,
      "learning_rate": 2.0931207077173788e-05,
      "loss": 1.4407,
      "step": 6000
    },
    {
      "epoch": 0.6983240223463687,
      "eval_cosine_accuracy": 0.9492493271827698,
      "eval_loss": 2.457711935043335,
      "eval_runtime": 28.3495,
      "eval_samples_per_second": 333.621,
      "eval_steps_per_second": 1.305,
      "step": 6000
    },
    {
      "epoch": 0.7565176908752328,
      "grad_norm": 0.6600456237792969,
      "learning_rate": 2.2677220346874637e-05,
      "loss": 1.3692,
      "step": 6500
    },
    {
      "epoch": 0.7565176908752328,
      "eval_cosine_accuracy": 0.9438570737838745,
      "eval_loss": 2.48168683052063,
      "eval_runtime": 28.3758,
      "eval_samples_per_second": 333.313,
      "eval_steps_per_second": 1.304,
      "step": 6500
    },
    {
      "epoch": 0.8147113594040968,
      "grad_norm": 15.486236572265625,
      "learning_rate": 2.4419741590036085e-05,
      "loss": 1.2731,
      "step": 7000
    },
    {
      "epoch": 0.8147113594040968,
      "eval_cosine_accuracy": 0.943751335144043,
      "eval_loss": 2.5139832496643066,
      "eval_runtime": 28.0964,
      "eval_samples_per_second": 336.627,
      "eval_steps_per_second": 1.317,
      "step": 7000
    },
    {
      "epoch": 0.8729050279329609,
      "grad_norm": 13.570350646972656,
      "learning_rate": 2.6165754859736934e-05,
      "loss": 1.223,
      "step": 7500
    },
    {
      "epoch": 0.8729050279329609,
      "eval_cosine_accuracy": 0.9494607448577881,
      "eval_loss": 2.4431588649749756,
      "eval_runtime": 28.2933,
      "eval_samples_per_second": 334.285,
      "eval_steps_per_second": 1.308,
      "step": 7500
    },
    {
      "epoch": 0.931098696461825,
      "grad_norm": 4.8542633056640625,
      "learning_rate": 2.7911768129437783e-05,
      "loss": 1.1982,
      "step": 8000
    },
    {
      "epoch": 0.931098696461825,
      "eval_cosine_accuracy": 0.9420596361160278,
      "eval_loss": 2.5187907218933105,
      "eval_runtime": 28.0669,
      "eval_samples_per_second": 336.98,
      "eval_steps_per_second": 1.318,
      "step": 8000
    },
    {
      "epoch": 0.9892923649906891,
      "grad_norm": 36.649906158447266,
      "learning_rate": 2.9657781399138632e-05,
      "loss": 1.1693,
      "step": 8500
    },
    {
      "epoch": 0.9892923649906891,
      "eval_cosine_accuracy": 0.9469232559204102,
      "eval_loss": 2.4668424129486084,
      "eval_runtime": 28.3845,
      "eval_samples_per_second": 333.21,
      "eval_steps_per_second": 1.304,
      "step": 8500
    }
  ],
  "logging_steps": 500,
  "max_steps": 34368,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 4,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 0.0,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}