{
  "best_global_step": 500,
  "best_metric": 2.0163703955210246,
  "best_model_checkpoint": "/content/cot-split3/checkpoint-500",
  "epoch": 0.19577900465954032,
  "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": 5.410557520272195,
      "eval_entropy": 3.805819726348172,
      "eval_loss": 3.754805564880371,
      "eval_mean_token_accuracy": 0.3303780525619185,
      "eval_nll": 3.7503126904340807,
      "eval_num_tokens": 0.0,
      "eval_ppl": 42.534380014623586,
      "eval_runtime": 319.2646,
      "eval_samples_per_second": 568.835,
      "eval_steps_per_second": 8.889,
      "step": 0
    },
    {
      "entropy": 3.939532442367398,
      "epoch": 0.00039155800931908063,
      "grad_norm": 2.921875,
      "learning_rate": 0.0,
      "loss": 3.6805007457733154,
      "mean_token_accuracy": 0.33797585886722376,
      "num_tokens": 159295.0,
      "step": 1
    },
    {
      "entropy": 4.131688260781753,
      "epoch": 0.009788950232977016,
      "grad_norm": 1.828125,
      "learning_rate": 9.375e-05,
      "loss": 3.7089433670043945,
      "mean_token_accuracy": 0.32488122625965454,
      "num_tokens": 3875056.0,
      "step": 25
    },
    {
      "entropy": 4.03907456745432,
      "epoch": 0.01957790046595403,
      "grad_norm": 0.87109375,
      "learning_rate": 0.00019140625,
      "loss": 3.2840447998046876,
      "mean_token_accuracy": 0.3501161635947849,
      "num_tokens": 7796334.0,
      "step": 50
    },
    {
      "entropy": 3.718109237212473,
      "epoch": 0.02936685069893105,
      "grad_norm": 0.875,
      "learning_rate": 0.0002890625,
      "loss": 2.907643127441406,
      "mean_token_accuracy": 0.37667007823156323,
      "num_tokens": 11688138.0,
      "step": 75
    },
    {
      "entropy": 3.4646144502832286,
      "epoch": 0.03915580093190806,
      "grad_norm": 0.71484375,
      "learning_rate": 0.00038671875,
      "loss": 2.5448974609375,
      "mean_token_accuracy": 0.41452789283904995,
      "num_tokens": 15554877.0,
      "step": 100
    },
    {
      "entropy": 3.2713017840492147,
      "epoch": 0.04894475116488508,
      "grad_norm": 0.71484375,
      "learning_rate": 0.000484375,
      "loss": 2.4108335876464846,
      "mean_token_accuracy": 0.4332291316341289,
      "num_tokens": 19437565.0,
      "step": 125
    },
    {
      "entropy": 3.177977246980637,
      "epoch": 0.0587337013978621,
      "grad_norm": 0.6796875,
      "learning_rate": 0.0004999075642455791,
      "loss": 2.35259765625,
      "mean_token_accuracy": 0.4402555277869966,
      "num_tokens": 23324888.0,
      "step": 150
    },
    {
      "entropy": 3.1149309732260146,
      "epoch": 0.06852265163083911,
      "grad_norm": 0.6171875,
      "learning_rate": 0.0004995565798713207,
      "loss": 2.2896893310546873,
      "mean_token_accuracy": 0.44857320203325,
      "num_tokens": 27225978.0,
      "step": 175
    },
    {
      "entropy": 3.0429316923380134,
      "epoch": 0.07831160186381612,
      "grad_norm": 0.7109375,
      "learning_rate": 0.0004989440608306114,
      "loss": 2.2477577209472654,
      "mean_token_accuracy": 0.45537929991850945,
      "num_tokens": 31124891.0,
      "step": 200
    },
    {
      "entropy": 3.000580134712211,
      "epoch": 0.08810055209679314,
      "grad_norm": 0.73046875,
      "learning_rate": 0.0004980706490418613,
      "loss": 2.2050234985351564,
      "mean_token_accuracy": 0.4608995247476681,
      "num_tokens": 35005557.0,
      "step": 225
    },
    {
      "entropy": 2.972177586040443,
      "epoch": 0.09788950232977016,
      "grad_norm": 0.64453125,
      "learning_rate": 0.0004969372598384225,
      "loss": 2.1969384765625,
      "mean_token_accuracy": 0.46284280083339074,
      "num_tokens": 38902413.0,
      "step": 250
    },
    {
      "epoch": 0.09788950232977016,
      "eval_bpd": 3.072745569254064,
      "eval_entropy": 2.8243270517127983,
      "eval_loss": 2.126833915710449,
      "eval_mean_token_accuracy": 0.4728344847516112,
      "eval_nll": 2.1298649279065187,
      "eval_num_tokens": 38902413.0,
      "eval_ppl": 8.41373027452248,
      "eval_runtime": 318.4023,
      "eval_samples_per_second": 570.376,
      "eval_steps_per_second": 8.913,
      "step": 250
    },
    {
      "entropy": 2.9943911642269376,
      "epoch": 0.10767845256274718,
      "grad_norm": 0.6328125,
      "learning_rate": 0.0004955450810093227,
      "loss": 2.189352722167969,
      "mean_token_accuracy": 0.4613144874232056,
      "num_tokens": 42787954.0,
      "step": 275
    },
    {
      "entropy": 2.9854789787582336,
      "epoch": 0.1174674027957242,
      "grad_norm": 0.64453125,
      "learning_rate": 0.0004938955715544644,
      "loss": 2.197265625,
      "mean_token_accuracy": 0.4586064115776618,
      "num_tokens": 46649640.0,
      "step": 300
    },
    {
      "entropy": 2.967871895537734,
      "epoch": 0.1272563530287012,
      "grad_norm": 0.77734375,
      "learning_rate": 0.0004919904601555969,
      "loss": 2.1636979675292967,
      "mean_token_accuracy": 0.46279846221708726,
      "num_tokens": 50562181.0,
      "step": 325
    },
    {
      "entropy": 2.9346317600010297,
      "epoch": 0.13704530326167821,
      "grad_norm": 0.66796875,
      "learning_rate": 0.0004898317433646626,
      "loss": 2.1643003845214843,
      "mean_token_accuracy": 0.4652492077421426,
      "num_tokens": 54435712.0,
      "step": 350
    },
    {
      "entropy": 2.8769926588271595,
      "epoch": 0.14683425349465523,
      "grad_norm": 0.81640625,
      "learning_rate": 0.00048742168351141525,
      "loss": 2.1070077514648435,
      "mean_token_accuracy": 0.4735955714097171,
      "num_tokens": 58331515.0,
      "step": 375
    },
    {
      "entropy": 2.9047266886544443,
      "epoch": 0.15662320372763225,
      "grad_norm": 0.58984375,
      "learning_rate": 0.00048476280633250404,
      "loss": 2.1295965576171874,
      "mean_token_accuracy": 0.4706730742865917,
      "num_tokens": 62245696.0,
      "step": 400
    },
    {
      "entropy": 2.8752667624718105,
      "epoch": 0.16641215396060927,
      "grad_norm": 0.58203125,
      "learning_rate": 0.00048185789832450813,
      "loss": 2.11253173828125,
      "mean_token_accuracy": 0.4743365019853278,
      "num_tokens": 66145585.0,
      "step": 425
    },
    {
      "entropy": 2.8552429066577076,
      "epoch": 0.17620110419358628,
      "grad_norm": 0.59375,
      "learning_rate": 0.00047871000382369524,
      "loss": 2.0846556091308592,
      "mean_token_accuracy": 0.47697299871989524,
      "num_tokens": 70063214.0,
      "step": 450
    },
    {
      "entropy": 2.8341120851783472,
      "epoch": 0.1859900544265633,
      "grad_norm": 0.58203125,
      "learning_rate": 0.0004753224218155648,
      "loss": 2.068045349121094,
      "mean_token_accuracy": 0.47787787454415265,
      "num_tokens": 73984544.0,
      "step": 475
    },
    {
      "entropy": 2.8375192761847376,
      "epoch": 0.19577900465954032,
      "grad_norm": 0.61328125,
      "learning_rate": 0.0004716987024775197,
      "loss": 2.0847515869140625,
      "mean_token_accuracy": 0.47784874690977935,
      "num_tokens": 77862885.0,
      "step": 500
    },
    {
      "epoch": 0.19577900465954032,
      "eval_bpd": 2.9090075702135,
      "eval_entropy": 2.6186967416048157,
      "eval_loss": 2.013190507888794,
      "eval_mean_token_accuracy": 0.48935668935428794,
      "eval_nll": 2.0163703955210246,
      "eval_num_tokens": 77862885.0,
      "eval_ppl": 7.5110133907515495,
      "eval_runtime": 318.4358,
      "eval_samples_per_second": 570.316,
      "eval_steps_per_second": 8.912,
      "step": 500
    }
  ],
  "logging_steps": 25,
  "max_steps": 2554,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "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": 2.525280992403456e+17,
  "train_batch_size": 64,
  "trial_name": null,
  "trial_params": null
}