{
  "best_global_step": 2500,
  "best_metric": 0.2838171720504761,
  "best_model_checkpoint": "output/checkpoint-2500",
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 2646,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.11344299489506524,
      "grad_norm": 0.6828517317771912,
      "learning_rate": 7.471698113207547e-05,
      "loss": 1.2849,
      "mean_token_accuracy": 0.7202034851908684,
      "num_tokens": 1109379.0,
      "step": 100
    },
    {
      "epoch": 0.22688598979013047,
      "grad_norm": 0.7256981134414673,
      "learning_rate": 0.00015018867924528303,
      "loss": 1.0128,
      "mean_token_accuracy": 0.7610476166009903,
      "num_tokens": 2235350.0,
      "step": 200
    },
    {
      "epoch": 0.3403289846851957,
      "grad_norm": 0.5375235080718994,
      "learning_rate": 0.0001998993912029063,
      "loss": 0.899,
      "mean_token_accuracy": 0.7830679216980934,
      "num_tokens": 3357507.0,
      "step": 300
    },
    {
      "epoch": 0.45377197958026094,
      "grad_norm": 0.5699286460876465,
      "learning_rate": 0.00019844106172649113,
      "loss": 0.802,
      "mean_token_accuracy": 0.8053381592035294,
      "num_tokens": 4495916.0,
      "step": 400
    },
    {
      "epoch": 0.5672149744753261,
      "grad_norm": 0.6252707242965698,
      "learning_rate": 0.00019527142855069377,
      "loss": 0.6939,
      "mean_token_accuracy": 0.8303170916438103,
      "num_tokens": 5638963.0,
      "step": 500
    },
    {
      "epoch": 0.5672149744753261,
      "eval_loss": 0.6935457587242126,
      "eval_mean_token_accuracy": 0.8294496539295936,
      "eval_num_tokens": 5638963.0,
      "eval_runtime": 276.8401,
      "eval_samples_per_second": 5.664,
      "eval_steps_per_second": 0.708,
      "step": 500
    },
    {
      "epoch": 0.6806579693703914,
      "grad_norm": 0.482766717672348,
      "learning_rate": 0.00019044559271652598,
      "loss": 0.6269,
      "mean_token_accuracy": 0.8460214346647262,
      "num_tokens": 6771799.0,
      "step": 600
    },
    {
      "epoch": 0.7941009642654566,
      "grad_norm": 0.6250064373016357,
      "learning_rate": 0.00018404744676405976,
      "loss": 0.5745,
      "mean_token_accuracy": 0.8582470047473908,
      "num_tokens": 7896784.0,
      "step": 700
    },
    {
      "epoch": 0.9075439591605219,
      "grad_norm": 0.696171760559082,
      "learning_rate": 0.00017618821634086138,
      "loss": 0.5008,
      "mean_token_accuracy": 0.8766051492094994,
      "num_tokens": 9036185.0,
      "step": 800
    },
    {
      "epoch": 1.0204197390811118,
      "grad_norm": 0.2615467309951782,
      "learning_rate": 0.00016700452665049568,
      "loss": 0.4528,
      "mean_token_accuracy": 0.8879575738355742,
      "num_tokens": 10264626.0,
      "step": 900
    },
    {
      "epoch": 1.1338627339761769,
      "grad_norm": 0.3127917945384979,
      "learning_rate": 0.00015665602735404375,
      "loss": 0.3348,
      "mean_token_accuracy": 0.9150179827213287,
      "num_tokens": 11398500.0,
      "step": 1000
    },
    {
      "epoch": 1.1338627339761769,
      "eval_loss": 0.44535523653030396,
      "eval_mean_token_accuracy": 0.8932840066904925,
      "eval_num_tokens": 11398500.0,
      "eval_runtime": 276.2931,
      "eval_samples_per_second": 5.675,
      "eval_steps_per_second": 0.709,
      "step": 1000
    },
    {
      "epoch": 1.2473057288712421,
      "grad_norm": 0.2831040620803833,
      "learning_rate": 0.00014532261721344984,
      "loss": 0.3248,
      "mean_token_accuracy": 0.9180563706159591,
      "num_tokens": 12511387.0,
      "step": 1100
    },
    {
      "epoch": 1.3607487237663074,
      "grad_norm": 0.2554262578487396,
      "learning_rate": 0.00013320131672361777,
      "loss": 0.3089,
      "mean_token_accuracy": 0.9220700481534004,
      "num_tokens": 13646331.0,
      "step": 1200
    },
    {
      "epoch": 1.4741917186613727,
      "grad_norm": 0.3375890851020813,
      "learning_rate": 0.00012050284309955694,
      "loss": 0.2894,
      "mean_token_accuracy": 0.9279368036985397,
      "num_tokens": 14794369.0,
      "step": 1300
    },
    {
      "epoch": 1.587634713556438,
      "grad_norm": 0.2744344472885132,
      "learning_rate": 0.00010744794715915005,
      "loss": 0.2748,
      "mean_token_accuracy": 0.9307059472799302,
      "num_tokens": 15925261.0,
      "step": 1400
    },
    {
      "epoch": 1.701077708451503,
      "grad_norm": 0.27422216534614563,
      "learning_rate": 9.4263575781332e-05,
      "loss": 0.2563,
      "mean_token_accuracy": 0.9356163629889488,
      "num_tokens": 17065748.0,
      "step": 1500
    },
    {
      "epoch": 1.701077708451503,
      "eval_loss": 0.3396897614002228,
      "eval_mean_token_accuracy": 0.9201750569805807,
      "eval_num_tokens": 17065748.0,
      "eval_runtime": 275.4276,
      "eval_samples_per_second": 5.693,
      "eval_steps_per_second": 0.712,
      "step": 1500
    },
    {
      "epoch": 1.8145207033465685,
      "grad_norm": 0.2150234878063202,
      "learning_rate": 8.117892665167406e-05,
      "loss": 0.238,
      "mean_token_accuracy": 0.9404442739486695,
      "num_tokens": 18190015.0,
      "step": 1600
    },
    {
      "epoch": 1.9279636982416335,
      "grad_norm": 0.2983264625072479,
      "learning_rate": 6.842146387984734e-05,
      "loss": 0.219,
      "mean_token_accuracy": 0.9447547772526741,
      "num_tokens": 19310575.0,
      "step": 1700
    },
    {
      "epoch": 2.0408394781622237,
      "grad_norm": 0.33862537145614624,
      "learning_rate": 5.621296375364546e-05,
      "loss": 0.2008,
      "mean_token_accuracy": 0.9498060237822221,
      "num_tokens": 20446504.0,
      "step": 1800
    },
    {
      "epoch": 2.1542824730572887,
      "grad_norm": 0.43845516443252563,
      "learning_rate": 4.4765659370349725e-05,
      "loss": 0.1256,
      "mean_token_accuracy": 0.9682124453783035,
      "num_tokens": 21558096.0,
      "step": 1900
    },
    {
      "epoch": 2.2677254679523537,
      "grad_norm": 0.4173540472984314,
      "learning_rate": 3.427855116733367e-05,
      "loss": 0.1206,
      "mean_token_accuracy": 0.9689858189225197,
      "num_tokens": 22703794.0,
      "step": 2000
    },
    {
      "epoch": 2.2677254679523537,
      "eval_loss": 0.2952539622783661,
      "eval_mean_token_accuracy": 0.9342006676051081,
      "eval_num_tokens": 22703794.0,
      "eval_runtime": 275.9272,
      "eval_samples_per_second": 5.683,
      "eval_steps_per_second": 0.71,
      "step": 2000
    },
    {
      "epoch": 2.3811684628474192,
      "grad_norm": 0.4896949529647827,
      "learning_rate": 2.493394748980298e-05,
      "loss": 0.1256,
      "mean_token_accuracy": 0.9678885874152183,
      "num_tokens": 23827405.0,
      "step": 2100
    },
    {
      "epoch": 2.4946114577424843,
      "grad_norm": 0.3831748068332672,
      "learning_rate": 1.6894295334591592e-05,
      "loss": 0.1123,
      "mean_token_accuracy": 0.9710856053233147,
      "num_tokens": 24940124.0,
      "step": 2200
    },
    {
      "epoch": 2.6080544526375498,
      "grad_norm": 0.4638579487800598,
      "learning_rate": 1.0299356364502255e-05,
      "loss": 0.1083,
      "mean_token_accuracy": 0.9723807728290558,
      "num_tokens": 26077789.0,
      "step": 2300
    },
    {
      "epoch": 2.721497447532615,
      "grad_norm": 0.4337230920791626,
      "learning_rate": 5.263777285480376e-06,
      "loss": 0.106,
      "mean_token_accuracy": 0.9729050415754318,
      "num_tokens": 27217807.0,
      "step": 2400
    },
    {
      "epoch": 2.83494044242768,
      "grad_norm": 0.3685455024242401,
      "learning_rate": 1.8750968232906785e-06,
      "loss": 0.1088,
      "mean_token_accuracy": 0.972323155105114,
      "num_tokens": 28349123.0,
      "step": 2500
    },
    {
      "epoch": 2.83494044242768,
      "eval_loss": 0.2838171720504761,
      "eval_mean_token_accuracy": 0.9382532765062488,
      "eval_num_tokens": 28349123.0,
      "eval_runtime": 275.4684,
      "eval_samples_per_second": 5.692,
      "eval_steps_per_second": 0.712,
      "step": 2500
    },
    {
      "epoch": 2.9483834373227453,
      "grad_norm": 0.43383386731147766,
      "learning_rate": 1.9222394650038056e-07,
      "loss": 0.1001,
      "mean_token_accuracy": 0.9745477658510208,
      "num_tokens": 29494601.0,
      "step": 2600
    }
  ],
  "logging_steps": 100,
  "max_steps": 2646,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "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": 1.3978374276266435e+18,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}