{
  "best_metric": 1.0532242059707642,
  "best_model_checkpoint": "outputs/checkpoint-648",
  "epoch": 19.865047233468285,
  "eval_steps": 500,
  "global_step": 920,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.4318488529014845,
      "grad_norm": 0.8586685061454773,
      "learning_rate": 6e-06,
      "loss": 2.3406,
      "step": 20
    },
    {
      "epoch": 0.863697705802969,
      "grad_norm": 0.6262326240539551,
      "learning_rate": 1.2e-05,
      "loss": 2.2977,
      "step": 40
    },
    {
      "epoch": 0.9932523616734144,
      "eval_loss": 2.046743154525757,
      "eval_runtime": 11.8178,
      "eval_samples_per_second": 31.478,
      "eval_steps_per_second": 3.977,
      "step": 46
    },
    {
      "epoch": 1.2955465587044535,
      "grad_norm": 0.5441118478775024,
      "learning_rate": 1.8e-05,
      "loss": 2.1983,
      "step": 60
    },
    {
      "epoch": 1.7273954116059378,
      "grad_norm": 0.5027536153793335,
      "learning_rate": 2.4e-05,
      "loss": 2.0126,
      "step": 80
    },
    {
      "epoch": 1.9865047233468287,
      "eval_loss": 1.756650686264038,
      "eval_runtime": 11.8327,
      "eval_samples_per_second": 31.438,
      "eval_steps_per_second": 3.972,
      "step": 92
    },
    {
      "epoch": 2.1592442645074224,
      "grad_norm": 0.5985815525054932,
      "learning_rate": 3e-05,
      "loss": 1.9539,
      "step": 100
    },
    {
      "epoch": 2.591093117408907,
      "grad_norm": 0.5876953601837158,
      "learning_rate": 2.9955987017756105e-05,
      "loss": 1.8487,
      "step": 120
    },
    {
      "epoch": 2.979757085020243,
      "eval_loss": 1.5829453468322754,
      "eval_runtime": 11.8342,
      "eval_samples_per_second": 31.434,
      "eval_steps_per_second": 3.972,
      "step": 138
    },
    {
      "epoch": 3.0229419703103915,
      "grad_norm": 0.9397473931312561,
      "learning_rate": 2.982420635670523e-05,
      "loss": 1.838,
      "step": 140
    },
    {
      "epoch": 3.454790823211876,
      "grad_norm": 0.9532083868980408,
      "learning_rate": 2.9605431358166687e-05,
      "loss": 1.7307,
      "step": 160
    },
    {
      "epoch": 3.8866396761133606,
      "grad_norm": 0.9994444847106934,
      "learning_rate": 2.9300945880823957e-05,
      "loss": 1.6956,
      "step": 180
    },
    {
      "epoch": 3.9946018893387314,
      "eval_loss": 1.4417527914047241,
      "eval_runtime": 11.8388,
      "eval_samples_per_second": 31.422,
      "eval_steps_per_second": 3.97,
      "step": 185
    },
    {
      "epoch": 4.318488529014845,
      "grad_norm": 1.2850545644760132,
      "learning_rate": 2.8912536766531424e-05,
      "loss": 1.6058,
      "step": 200
    },
    {
      "epoch": 4.75033738191633,
      "grad_norm": 1.5529141426086426,
      "learning_rate": 2.8442483354415835e-05,
      "loss": 1.5981,
      "step": 220
    },
    {
      "epoch": 4.987854251012146,
      "eval_loss": 1.3265655040740967,
      "eval_runtime": 11.8379,
      "eval_samples_per_second": 31.424,
      "eval_steps_per_second": 3.97,
      "step": 231
    },
    {
      "epoch": 5.182186234817814,
      "grad_norm": 1.335929036140442,
      "learning_rate": 2.789354410480802e-05,
      "loss": 1.5531,
      "step": 240
    },
    {
      "epoch": 5.614035087719298,
      "grad_norm": 1.6453096866607666,
      "learning_rate": 2.7268940411500768e-05,
      "loss": 1.4762,
      "step": 260
    },
    {
      "epoch": 5.98110661268556,
      "eval_loss": 1.2389419078826904,
      "eval_runtime": 11.8303,
      "eval_samples_per_second": 31.445,
      "eval_steps_per_second": 3.973,
      "step": 277
    },
    {
      "epoch": 6.045883940620783,
      "grad_norm": 1.6546735763549805,
      "learning_rate": 2.6572337697329145e-05,
      "loss": 1.4335,
      "step": 280
    },
    {
      "epoch": 6.477732793522267,
      "grad_norm": 1.6672505140304565,
      "learning_rate": 2.5807823904011803e-05,
      "loss": 1.3456,
      "step": 300
    },
    {
      "epoch": 6.909581646423752,
      "grad_norm": 1.872035264968872,
      "learning_rate": 2.497988550248348e-05,
      "loss": 1.3907,
      "step": 320
    },
    {
      "epoch": 6.995951417004049,
      "eval_loss": 1.1727797985076904,
      "eval_runtime": 11.8334,
      "eval_samples_per_second": 31.436,
      "eval_steps_per_second": 3.972,
      "step": 324
    },
    {
      "epoch": 7.341430499325236,
      "grad_norm": 2.2928590774536133,
      "learning_rate": 2.4093381164499572e-05,
      "loss": 1.2972,
      "step": 340
    },
    {
      "epoch": 7.77327935222672,
      "grad_norm": 2.046597957611084,
      "learning_rate": 2.315351325001832e-05,
      "loss": 1.3121,
      "step": 360
    },
    {
      "epoch": 7.989203778677463,
      "eval_loss": 1.1317445039749146,
      "eval_runtime": 11.8354,
      "eval_samples_per_second": 31.431,
      "eval_steps_per_second": 3.971,
      "step": 370
    },
    {
      "epoch": 8.205128205128204,
      "grad_norm": 2.3657939434051514,
      "learning_rate": 2.2165797277683945e-05,
      "loss": 1.2587,
      "step": 380
    },
    {
      "epoch": 8.63697705802969,
      "grad_norm": 2.6628217697143555,
      "learning_rate": 2.11360295575701e-05,
      "loss": 1.2259,
      "step": 400
    },
    {
      "epoch": 8.982456140350877,
      "eval_loss": 1.089838981628418,
      "eval_runtime": 11.8331,
      "eval_samples_per_second": 31.437,
      "eval_steps_per_second": 3.972,
      "step": 416
    },
    {
      "epoch": 9.068825910931174,
      "grad_norm": 2.126490354537964,
      "learning_rate": 2.007025317612754e-05,
      "loss": 1.2231,
      "step": 420
    },
    {
      "epoch": 9.50067476383266,
      "grad_norm": 2.33013653755188,
      "learning_rate": 1.897472253294993e-05,
      "loss": 1.1601,
      "step": 440
    },
    {
      "epoch": 9.932523616734143,
      "grad_norm": 2.1320955753326416,
      "learning_rate": 1.7855866637470027e-05,
      "loss": 1.1858,
      "step": 460
    },
    {
      "epoch": 9.997300944669366,
      "eval_loss": 1.0753756761550903,
      "eval_runtime": 11.8356,
      "eval_samples_per_second": 31.431,
      "eval_steps_per_second": 3.971,
      "step": 463
    },
    {
      "epoch": 10.364372469635628,
      "grad_norm": 2.7270896434783936,
      "learning_rate": 1.6720251380976008e-05,
      "loss": 1.1264,
      "step": 480
    },
    {
      "epoch": 10.796221322537113,
      "grad_norm": 2.817499876022339,
      "learning_rate": 1.557454100535053e-05,
      "loss": 1.0953,
      "step": 500
    },
    {
      "epoch": 10.99055330634278,
      "eval_loss": 1.0628381967544556,
      "eval_runtime": 11.8346,
      "eval_samples_per_second": 31.433,
      "eval_steps_per_second": 3.971,
      "step": 509
    },
    {
      "epoch": 11.228070175438596,
      "grad_norm": 2.4983890056610107,
      "learning_rate": 1.442545899464947e-05,
      "loss": 1.1448,
      "step": 520
    },
    {
      "epoch": 11.65991902834008,
      "grad_norm": 3.187267303466797,
      "learning_rate": 1.3279748619023994e-05,
      "loss": 1.0969,
      "step": 540
    },
    {
      "epoch": 11.983805668016194,
      "eval_loss": 1.0652039051055908,
      "eval_runtime": 11.8314,
      "eval_samples_per_second": 31.442,
      "eval_steps_per_second": 3.972,
      "step": 555
    },
    {
      "epoch": 12.091767881241566,
      "grad_norm": 2.7236690521240234,
      "learning_rate": 1.2144133362529972e-05,
      "loss": 1.0659,
      "step": 560
    },
    {
      "epoch": 12.523616734143049,
      "grad_norm": 2.291569232940674,
      "learning_rate": 1.1025277467050077e-05,
      "loss": 1.0509,
      "step": 580
    },
    {
      "epoch": 12.955465587044534,
      "grad_norm": 3.7598021030426025,
      "learning_rate": 9.929746823872461e-06,
      "loss": 1.0398,
      "step": 600
    },
    {
      "epoch": 12.998650472334683,
      "eval_loss": 1.0660185813903809,
      "eval_runtime": 11.8359,
      "eval_samples_per_second": 31.43,
      "eval_steps_per_second": 3.971,
      "step": 602
    },
    {
      "epoch": 13.387314439946019,
      "grad_norm": 2.636166572570801,
      "learning_rate": 8.863970442429903e-06,
      "loss": 1.0089,
      "step": 620
    },
    {
      "epoch": 13.819163292847504,
      "grad_norm": 2.8337323665618896,
      "learning_rate": 7.834202722316054e-06,
      "loss": 1.0692,
      "step": 640
    },
    {
      "epoch": 13.991902834008098,
      "eval_loss": 1.0532242059707642,
      "eval_runtime": 11.8398,
      "eval_samples_per_second": 31.419,
      "eval_steps_per_second": 3.97,
      "step": 648
    },
    {
      "epoch": 14.251012145748987,
      "grad_norm": 3.053009033203125,
      "learning_rate": 6.846486749981685e-06,
      "loss": 1.0366,
      "step": 660
    },
    {
      "epoch": 14.682860998650472,
      "grad_norm": 2.598278522491455,
      "learning_rate": 5.906618835500434e-06,
      "loss": 0.9994,
      "step": 680
    },
    {
      "epoch": 14.98515519568151,
      "eval_loss": 1.0662660598754883,
      "eval_runtime": 11.8811,
      "eval_samples_per_second": 31.31,
      "eval_steps_per_second": 3.956,
      "step": 694
    },
    {
      "epoch": 15.114709851551957,
      "grad_norm": 3.115391492843628,
      "learning_rate": 5.0201144975165215e-06,
      "loss": 1.0248,
      "step": 700
    },
    {
      "epoch": 15.54655870445344,
      "grad_norm": 2.9552927017211914,
      "learning_rate": 4.192176095988196e-06,
      "loss": 0.9757,
      "step": 720
    },
    {
      "epoch": 15.978407557354926,
      "grad_norm": 2.836552143096924,
      "learning_rate": 3.4276623026708552e-06,
      "loss": 1.0462,
      "step": 740
    },
    {
      "epoch": 16.0,
      "eval_loss": 1.0600887537002563,
      "eval_runtime": 11.8762,
      "eval_samples_per_second": 31.323,
      "eval_steps_per_second": 3.957,
      "step": 741
    },
    {
      "epoch": 16.41025641025641,
      "grad_norm": 2.7339868545532227,
      "learning_rate": 2.7310595884992356e-06,
      "loss": 0.956,
      "step": 760
    },
    {
      "epoch": 16.842105263157894,
      "grad_norm": 2.860797882080078,
      "learning_rate": 2.1064558951919854e-06,
      "loss": 1.0121,
      "step": 780
    },
    {
      "epoch": 16.993252361673413,
      "eval_loss": 1.067551851272583,
      "eval_runtime": 11.9433,
      "eval_samples_per_second": 31.147,
      "eval_steps_per_second": 3.935,
      "step": 787
    },
    {
      "epoch": 17.27395411605938,
      "grad_norm": 3.2689273357391357,
      "learning_rate": 1.5575166455841678e-06,
      "loss": 1.0109,
      "step": 800
    },
    {
      "epoch": 17.705802968960864,
      "grad_norm": 2.9255502223968506,
      "learning_rate": 1.0874632334685808e-06,
      "loss": 0.926,
      "step": 820
    },
    {
      "epoch": 17.98650472334683,
      "eval_loss": 1.0671662092208862,
      "eval_runtime": 11.9354,
      "eval_samples_per_second": 31.168,
      "eval_steps_per_second": 3.938,
      "step": 833
    },
    {
      "epoch": 18.13765182186235,
      "grad_norm": 2.799520492553711,
      "learning_rate": 6.990541191760419e-07,
      "loss": 1.015,
      "step": 840
    },
    {
      "epoch": 18.569500674763834,
      "grad_norm": 3.2285776138305664,
      "learning_rate": 3.945686418333155e-07,
      "loss": 0.9512,
      "step": 860
    },
    {
      "epoch": 18.979757085020243,
      "eval_loss": 1.0708719491958618,
      "eval_runtime": 11.9397,
      "eval_samples_per_second": 31.157,
      "eval_steps_per_second": 3.936,
      "step": 879
    },
    {
      "epoch": 19.00134952766532,
      "grad_norm": 2.990569591522217,
      "learning_rate": 1.7579364329477376e-07,
      "loss": 0.9571,
      "step": 880
    },
    {
      "epoch": 19.4331983805668,
      "grad_norm": 3.1629343032836914,
      "learning_rate": 4.401298224389338e-08,
      "loss": 0.9524,
      "step": 900
    },
    {
      "epoch": 19.865047233468285,
      "grad_norm": 3.00166392326355,
      "learning_rate": 0.0,
      "loss": 0.976,
      "step": 920
    },
    {
      "epoch": 19.865047233468285,
      "eval_loss": 1.0710477828979492,
      "eval_runtime": 11.9108,
      "eval_samples_per_second": 31.232,
      "eval_steps_per_second": 3.946,
      "step": 920
    }
  ],
  "logging_steps": 20,
  "max_steps": 920,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 20,
  "save_steps": 500,
  "total_flos": 8.174619616582042e+16,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}