{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 4.0,
  "eval_steps": 1000,
  "global_step": 15734,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.000254323499491353,
      "grad_norm": 6.461288928985596,
      "learning_rate": 0.0,
      "loss": 3.7722,
      "step": 1
    },
    {
      "epoch": 0.254323499491353,
      "grad_norm": 5.289657115936279,
      "learning_rate": 1.5244150559511698e-05,
      "loss": 2.8712,
      "step": 1000
    },
    {
      "epoch": 0.254323499491353,
      "eval_cosine_accuracy": 0.9444736838340759,
      "eval_loss": 0.5577611327171326,
      "eval_runtime": 35.3135,
      "eval_samples_per_second": 269.274,
      "eval_steps_per_second": 1.076,
      "step": 1000
    },
    {
      "epoch": 0.508646998982706,
      "grad_norm": 6.285799026489258,
      "learning_rate": 2.9944048830111903e-05,
      "loss": 1.8052,
      "step": 2000
    },
    {
      "epoch": 0.508646998982706,
      "eval_cosine_accuracy": 0.9498369693756104,
      "eval_loss": 0.5062179565429688,
      "eval_runtime": 34.9468,
      "eval_samples_per_second": 272.099,
      "eval_steps_per_second": 1.087,
      "step": 2000
    },
    {
      "epoch": 0.762970498474059,
      "grad_norm": 6.981422424316406,
      "learning_rate": 2.8250254323499494e-05,
      "loss": 1.2908,
      "step": 3000
    },
    {
      "epoch": 0.762970498474059,
      "eval_cosine_accuracy": 0.9575139284133911,
      "eval_loss": 0.4583439528942108,
      "eval_runtime": 34.7917,
      "eval_samples_per_second": 273.312,
      "eval_steps_per_second": 1.092,
      "step": 3000
    },
    {
      "epoch": 1.0172852058973056,
      "grad_norm": 6.195133209228516,
      "learning_rate": 2.6554764326890472e-05,
      "loss": 1.0201,
      "step": 4000
    },
    {
      "epoch": 1.0172852058973056,
      "eval_cosine_accuracy": 0.9621411561965942,
      "eval_loss": 0.4364590048789978,
      "eval_runtime": 35.4237,
      "eval_samples_per_second": 268.436,
      "eval_steps_per_second": 1.073,
      "step": 4000
    },
    {
      "epoch": 1.2714794102694458,
      "grad_norm": 6.555699825286865,
      "learning_rate": 2.4859274330281453e-05,
      "loss": 1.2224,
      "step": 5000
    },
    {
      "epoch": 1.2714794102694458,
      "eval_cosine_accuracy": 0.962666928768158,
      "eval_loss": 0.4258391261100769,
      "eval_runtime": 36.1657,
      "eval_samples_per_second": 262.928,
      "eval_steps_per_second": 1.051,
      "step": 5000
    },
    {
      "epoch": 1.5256736146415861,
      "grad_norm": 4.430365562438965,
      "learning_rate": 2.316378433367243e-05,
      "loss": 1.1341,
      "step": 6000
    },
    {
      "epoch": 1.5256736146415861,
      "eval_cosine_accuracy": 0.964454710483551,
      "eval_loss": 0.4155160188674927,
      "eval_runtime": 35.2226,
      "eval_samples_per_second": 269.969,
      "eval_steps_per_second": 1.079,
      "step": 6000
    },
    {
      "epoch": 1.7798678190137265,
      "grad_norm": 5.276082515716553,
      "learning_rate": 2.146998982706002e-05,
      "loss": 1.0798,
      "step": 7000
    },
    {
      "epoch": 1.7798678190137265,
      "eval_cosine_accuracy": 0.9657166600227356,
      "eval_loss": 0.42145127058029175,
      "eval_runtime": 35.2976,
      "eval_samples_per_second": 269.395,
      "eval_steps_per_second": 1.077,
      "step": 7000
    },
    {
      "epoch": 2.0340620233858666,
      "grad_norm": 4.906863689422607,
      "learning_rate": 1.9774499830451e-05,
      "loss": 1.0249,
      "step": 8000
    },
    {
      "epoch": 2.0340620233858666,
      "eval_cosine_accuracy": 0.9681354761123657,
      "eval_loss": 0.41128453612327576,
      "eval_runtime": 35.164,
      "eval_samples_per_second": 270.419,
      "eval_steps_per_second": 1.081,
      "step": 8000
    },
    {
      "epoch": 2.288256227758007,
      "grad_norm": 6.295294761657715,
      "learning_rate": 1.808070532383859e-05,
      "loss": 0.9443,
      "step": 9000
    },
    {
      "epoch": 2.288256227758007,
      "eval_cosine_accuracy": 0.9672941565513611,
      "eval_loss": 0.4052944779396057,
      "eval_runtime": 35.2065,
      "eval_samples_per_second": 270.092,
      "eval_steps_per_second": 1.079,
      "step": 9000
    },
    {
      "epoch": 2.5424504321301473,
      "grad_norm": 5.057905197143555,
      "learning_rate": 1.6386910817226178e-05,
      "loss": 0.9178,
      "step": 10000
    },
    {
      "epoch": 2.5424504321301473,
      "eval_cosine_accuracy": 0.968766450881958,
      "eval_loss": 0.39761021733283997,
      "eval_runtime": 35.1256,
      "eval_samples_per_second": 270.714,
      "eval_steps_per_second": 1.082,
      "step": 10000
    },
    {
      "epoch": 2.796644636502288,
      "grad_norm": 5.986538410186768,
      "learning_rate": 1.4691420820617159e-05,
      "loss": 0.8991,
      "step": 11000
    },
    {
      "epoch": 2.796644636502288,
      "eval_cosine_accuracy": 0.9703438878059387,
      "eval_loss": 0.40334755182266235,
      "eval_runtime": 36.7674,
      "eval_samples_per_second": 258.626,
      "eval_steps_per_second": 1.034,
      "step": 11000
    },
    {
      "epoch": 3.050838840874428,
      "grad_norm": 4.371334075927734,
      "learning_rate": 1.2995930824008138e-05,
      "loss": 0.8777,
      "step": 12000
    },
    {
      "epoch": 3.050838840874428,
      "eval_cosine_accuracy": 0.9698180556297302,
      "eval_loss": 0.4055646061897278,
      "eval_runtime": 35.5697,
      "eval_samples_per_second": 267.334,
      "eval_steps_per_second": 1.068,
      "step": 12000
    },
    {
      "epoch": 3.305033045246568,
      "grad_norm": 6.505381107330322,
      "learning_rate": 1.1303831807392337e-05,
      "loss": 0.8326,
      "step": 13000
    },
    {
      "epoch": 3.305033045246568,
      "eval_cosine_accuracy": 0.9692922234535217,
      "eval_loss": 0.40629708766937256,
      "eval_runtime": 35.3933,
      "eval_samples_per_second": 268.667,
      "eval_steps_per_second": 1.074,
      "step": 13000
    },
    {
      "epoch": 3.5592272496187087,
      "grad_norm": 4.900751113891602,
      "learning_rate": 9.608341810783316e-06,
      "loss": 0.8178,
      "step": 14000
    },
    {
      "epoch": 3.5592272496187087,
      "eval_cosine_accuracy": 0.9692922234535217,
      "eval_loss": 0.40360188484191895,
      "eval_runtime": 35.5577,
      "eval_samples_per_second": 267.424,
      "eval_steps_per_second": 1.069,
      "step": 14000
    },
    {
      "epoch": 3.813421453990849,
      "grad_norm": 4.314871788024902,
      "learning_rate": 7.912851814174296e-06,
      "loss": 0.8067,
      "step": 15000
    },
    {
      "epoch": 3.813421453990849,
      "eval_cosine_accuracy": 0.9695025682449341,
      "eval_loss": 0.3971245586872101,
      "eval_runtime": 35.4781,
      "eval_samples_per_second": 268.024,
      "eval_steps_per_second": 1.071,
      "step": 15000
    }
  ],
  "logging_steps": 1000,
  "max_steps": 19660,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 5,
  "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
}