{
  "best_global_step": 500,
  "best_metric": 2.009260514264372,
  "best_model_checkpoint": "/content/cot-split1/checkpoint-500",
  "epoch": 0.7832080200501254,
  "eval_steps": 250,
  "global_step": 500,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0,
      "eval_bpd": 5.449506205976702,
      "eval_entropy": 3.8259646681880883,
      "eval_loss": 3.7819511890411377,
      "eval_mean_token_accuracy": 0.3298281561563004,
      "eval_nll": 3.777309862116675,
      "eval_num_tokens": 0.0,
      "eval_ppl": 43.69832894461198,
      "eval_runtime": 318.4397,
      "eval_samples_per_second": 570.309,
      "eval_steps_per_second": 8.912,
      "step": 0
    },
    {
      "entropy": 3.936198543874438,
      "epoch": 0.0015664160401002505,
      "grad_norm": 3.5,
      "learning_rate": 0.0,
      "loss": 3.687645435333252,
      "mean_token_accuracy": 0.34163331851193124,
      "num_tokens": 154043.0,
      "step": 1
    },
    {
      "entropy": 4.1582471953791025,
      "epoch": 0.039160401002506263,
      "grad_norm": 1.8515625,
      "learning_rate": 9.375e-05,
      "loss": 3.716327985127767,
      "mean_token_accuracy": 0.32336124196593746,
      "num_tokens": 3937675.0,
      "step": 25
    },
    {
      "entropy": 4.004832741749853,
      "epoch": 0.07832080200501253,
      "grad_norm": 1.578125,
      "learning_rate": 0.00019140625,
      "loss": 3.245400085449219,
      "mean_token_accuracy": 0.3525532396162817,
      "num_tokens": 7842040.0,
      "step": 50
    },
    {
      "entropy": 3.7060696659038226,
      "epoch": 0.1174812030075188,
      "grad_norm": 1.1328125,
      "learning_rate": 0.0002890625,
      "loss": 2.899266357421875,
      "mean_token_accuracy": 0.3803508952459354,
      "num_tokens": 11723384.0,
      "step": 75
    },
    {
      "entropy": 3.483661951565247,
      "epoch": 0.15664160401002505,
      "grad_norm": 0.796875,
      "learning_rate": 0.00038671875,
      "loss": 2.5580235290527344,
      "mean_token_accuracy": 0.4106306236210145,
      "num_tokens": 15613476.0,
      "step": 100
    },
    {
      "entropy": 3.279625251379403,
      "epoch": 0.19580200501253134,
      "grad_norm": 0.71484375,
      "learning_rate": 0.000484375,
      "loss": 2.3971221923828123,
      "mean_token_accuracy": 0.43163137955735414,
      "num_tokens": 19537696.0,
      "step": 125
    },
    {
      "entropy": 3.1596573478657435,
      "epoch": 0.2349624060150376,
      "grad_norm": 0.640625,
      "learning_rate": 0.0004999075642455791,
      "loss": 2.332003479003906,
      "mean_token_accuracy": 0.44275144542816425,
      "num_tokens": 23462566.0,
      "step": 150
    },
    {
      "entropy": 3.104312514398783,
      "epoch": 0.2741228070175439,
      "grad_norm": 0.609375,
      "learning_rate": 0.0004995565798713207,
      "loss": 2.2806515502929687,
      "mean_token_accuracy": 0.44779701792125276,
      "num_tokens": 27367329.0,
      "step": 175
    },
    {
      "entropy": 3.062113077731651,
      "epoch": 0.3132832080200501,
      "grad_norm": 0.796875,
      "learning_rate": 0.0004989440608306114,
      "loss": 2.2435511779785156,
      "mean_token_accuracy": 0.45369474441778246,
      "num_tokens": 31290903.0,
      "step": 200
    },
    {
      "entropy": 3.007945648412728,
      "epoch": 0.3524436090225564,
      "grad_norm": 0.71875,
      "learning_rate": 0.0004980706490418613,
      "loss": 2.214092559814453,
      "mean_token_accuracy": 0.46001344086021506,
      "num_tokens": 35158603.0,
      "step": 225
    },
    {
      "entropy": 2.9779412397442524,
      "epoch": 0.3916040100250627,
      "grad_norm": 0.65625,
      "learning_rate": 0.0004969372598384225,
      "loss": 2.191622314453125,
      "mean_token_accuracy": 0.46235443734230136,
      "num_tokens": 39050018.0,
      "step": 250
    },
    {
      "epoch": 0.3916040100250627,
      "eval_bpd": 3.058715802573539,
      "eval_entropy": 2.8622500597497296,
      "eval_loss": 2.117133855819702,
      "eval_mean_token_accuracy": 0.4719805146180118,
      "eval_nll": 2.120140234687999,
      "eval_num_tokens": 39050018.0,
      "eval_ppl": 8.332305884076892,
      "eval_runtime": 317.9266,
      "eval_samples_per_second": 571.229,
      "eval_steps_per_second": 8.927,
      "step": 250
    },
    {
      "entropy": 2.91319560400022,
      "epoch": 0.4307644110275689,
      "grad_norm": 0.59375,
      "learning_rate": 0.0004955450810093227,
      "loss": 2.1354435729980468,
      "mean_token_accuracy": 0.46928864274436577,
      "num_tokens": 42938915.0,
      "step": 275
    },
    {
      "entropy": 2.974404757708524,
      "epoch": 0.4699248120300752,
      "grad_norm": 0.63671875,
      "learning_rate": 0.0004938955715544644,
      "loss": 2.180239562988281,
      "mean_token_accuracy": 0.4623402148856342,
      "num_tokens": 46830642.0,
      "step": 300
    },
    {
      "entropy": 2.947778592589299,
      "epoch": 0.5090852130325815,
      "grad_norm": 0.52734375,
      "learning_rate": 0.0004919904601555969,
      "loss": 2.137745666503906,
      "mean_token_accuracy": 0.4655975033913256,
      "num_tokens": 50738889.0,
      "step": 325
    },
    {
      "entropy": 2.912182858652311,
      "epoch": 0.5482456140350878,
      "grad_norm": 0.73046875,
      "learning_rate": 0.0004898317433646626,
      "loss": 2.134133148193359,
      "mean_token_accuracy": 0.47067078409410695,
      "num_tokens": 54590447.0,
      "step": 350
    },
    {
      "entropy": 2.8808906902937665,
      "epoch": 0.5874060150375939,
      "grad_norm": 0.71484375,
      "learning_rate": 0.00048742168351141525,
      "loss": 2.111539001464844,
      "mean_token_accuracy": 0.47415726155363475,
      "num_tokens": 58477393.0,
      "step": 375
    },
    {
      "entropy": 2.8902592806431113,
      "epoch": 0.6265664160401002,
      "grad_norm": 0.77734375,
      "learning_rate": 0.00048476280633250404,
      "loss": 2.1172853088378907,
      "mean_token_accuracy": 0.47171053962916015,
      "num_tokens": 62364053.0,
      "step": 400
    },
    {
      "entropy": 2.8472467625827407,
      "epoch": 0.6657268170426065,
      "grad_norm": 0.5546875,
      "learning_rate": 0.00048185789832450813,
      "loss": 2.079963836669922,
      "mean_token_accuracy": 0.47701584614870657,
      "num_tokens": 66254150.0,
      "step": 425
    },
    {
      "entropy": 2.849270481153217,
      "epoch": 0.7048872180451128,
      "grad_norm": 0.76171875,
      "learning_rate": 0.00047871000382369524,
      "loss": 2.076465911865234,
      "mean_token_accuracy": 0.477678835790845,
      "num_tokens": 70155017.0,
      "step": 450
    },
    {
      "entropy": 2.8533846163704104,
      "epoch": 0.7440476190476191,
      "grad_norm": 0.75,
      "learning_rate": 0.0004753224218155648,
      "loss": 2.076942901611328,
      "mean_token_accuracy": 0.4763377187304664,
      "num_tokens": 74017889.0,
      "step": 475
    },
    {
      "entropy": 2.833587274583335,
      "epoch": 0.7832080200501254,
      "grad_norm": 0.61328125,
      "learning_rate": 0.0004716987024775197,
      "loss": 2.0738652038574217,
      "mean_token_accuracy": 0.48083406578747806,
      "num_tokens": 77928779.0,
      "step": 500
    },
    {
      "epoch": 0.7832080200501254,
      "eval_bpd": 2.898750179783218,
      "eval_entropy": 2.6689496127647594,
      "eval_loss": 2.006093978881836,
      "eval_mean_token_accuracy": 0.4891084327887536,
      "eval_nll": 2.009260514264372,
      "eval_num_tokens": 77928779.0,
      "eval_ppl": 7.457800370714545,
      "eval_runtime": 317.8942,
      "eval_samples_per_second": 571.288,
      "eval_steps_per_second": 8.927,
      "step": 500
    }
  ],
  "logging_steps": 25,
  "max_steps": 2554,
  "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": 2.5466593463033856e+17,
  "train_batch_size": 64,
  "trial_name": null,
  "trial_params": null
}