{
  "best_metric": 23.6596,
  "best_model_checkpoint": "/local1/hfs/gs_stuff/ft-wmt14/checkpoint-100000",
  "epoch": 2.7777777777777777,
  "eval_steps": 10000,
  "global_step": 100000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.1388888888888889,
      "grad_norm": 1.066943645477295,
      "learning_rate": 0.000475,
      "loss": 1.9627,
      "step": 5000
    },
    {
      "epoch": 0.2777777777777778,
      "grad_norm": 0.9774492383003235,
      "learning_rate": 0.00045000000000000004,
      "loss": 1.7738,
      "step": 10000
    },
    {
      "epoch": 0.2777777777777778,
      "eval_bleu": 20.1598,
      "eval_gen_len": 28.1563,
      "eval_loss": 1.914583444595337,
      "eval_runtime": 241.8013,
      "eval_samples_per_second": 12.407,
      "eval_steps_per_second": 1.551,
      "step": 10000
    },
    {
      "epoch": 0.4166666666666667,
      "grad_norm": 1.4306731224060059,
      "learning_rate": 0.000425,
      "loss": 1.6951,
      "step": 15000
    },
    {
      "epoch": 0.5555555555555556,
      "grad_norm": 1.1782424449920654,
      "learning_rate": 0.0004,
      "loss": 1.6498,
      "step": 20000
    },
    {
      "epoch": 0.5555555555555556,
      "eval_bleu": 21.4167,
      "eval_gen_len": 27.853,
      "eval_loss": 1.855008840560913,
      "eval_runtime": 242.3949,
      "eval_samples_per_second": 12.376,
      "eval_steps_per_second": 1.547,
      "step": 20000
    },
    {
      "epoch": 0.6944444444444444,
      "grad_norm": 1.219376802444458,
      "learning_rate": 0.000375,
      "loss": 1.6172,
      "step": 25000
    },
    {
      "epoch": 0.8333333333333334,
      "grad_norm": 1.2735612392425537,
      "learning_rate": 0.00035,
      "loss": 1.5903,
      "step": 30000
    },
    {
      "epoch": 0.8333333333333334,
      "eval_bleu": 22.604,
      "eval_gen_len": 27.7613,
      "eval_loss": 1.8276705741882324,
      "eval_runtime": 240.5149,
      "eval_samples_per_second": 12.473,
      "eval_steps_per_second": 1.559,
      "step": 30000
    },
    {
      "epoch": 0.9722222222222222,
      "grad_norm": 1.0282609462738037,
      "learning_rate": 0.00032500000000000004,
      "loss": 1.5633,
      "step": 35000
    },
    {
      "epoch": 1.1111111111111112,
      "grad_norm": 1.406827688217163,
      "learning_rate": 0.0003,
      "loss": 1.5151,
      "step": 40000
    },
    {
      "epoch": 1.1111111111111112,
      "eval_bleu": 22.1273,
      "eval_gen_len": 27.3187,
      "eval_loss": 1.8127936124801636,
      "eval_runtime": 234.7049,
      "eval_samples_per_second": 12.782,
      "eval_steps_per_second": 1.598,
      "step": 40000
    },
    {
      "epoch": 1.25,
      "grad_norm": 1.174306035041809,
      "learning_rate": 0.000275,
      "loss": 1.5004,
      "step": 45000
    },
    {
      "epoch": 1.3888888888888888,
      "grad_norm": 1.5665515661239624,
      "learning_rate": 0.00025,
      "loss": 1.4866,
      "step": 50000
    },
    {
      "epoch": 1.3888888888888888,
      "eval_bleu": 22.8295,
      "eval_gen_len": 27.419,
      "eval_loss": 1.7999275922775269,
      "eval_runtime": 233.8115,
      "eval_samples_per_second": 12.831,
      "eval_steps_per_second": 1.604,
      "step": 50000
    },
    {
      "epoch": 1.5277777777777777,
      "grad_norm": 1.1425319910049438,
      "learning_rate": 0.00022500000000000002,
      "loss": 1.4799,
      "step": 55000
    },
    {
      "epoch": 1.6666666666666665,
      "grad_norm": 1.123904824256897,
      "learning_rate": 0.0002,
      "loss": 1.4696,
      "step": 60000
    },
    {
      "epoch": 1.6666666666666665,
      "eval_bleu": 22.9923,
      "eval_gen_len": 27.7387,
      "eval_loss": 1.780959963798523,
      "eval_runtime": 240.0938,
      "eval_samples_per_second": 12.495,
      "eval_steps_per_second": 1.562,
      "step": 60000
    },
    {
      "epoch": 1.8055555555555556,
      "grad_norm": 1.4292243719100952,
      "learning_rate": 0.000175,
      "loss": 1.4613,
      "step": 65000
    },
    {
      "epoch": 1.9444444444444444,
      "grad_norm": 1.1662226915359497,
      "learning_rate": 0.00015,
      "loss": 1.4508,
      "step": 70000
    },
    {
      "epoch": 1.9444444444444444,
      "eval_bleu": 23.1046,
      "eval_gen_len": 27.7057,
      "eval_loss": 1.7654317617416382,
      "eval_runtime": 236.6367,
      "eval_samples_per_second": 12.678,
      "eval_steps_per_second": 1.585,
      "step": 70000
    },
    {
      "epoch": 2.0833333333333335,
      "grad_norm": 0.9245423674583435,
      "learning_rate": 0.000125,
      "loss": 1.4235,
      "step": 75000
    },
    {
      "epoch": 2.2222222222222223,
      "grad_norm": 1.2502944469451904,
      "learning_rate": 0.0001,
      "loss": 1.4053,
      "step": 80000
    },
    {
      "epoch": 2.2222222222222223,
      "eval_bleu": 23.5079,
      "eval_gen_len": 27.643,
      "eval_loss": 1.758699655532837,
      "eval_runtime": 237.5663,
      "eval_samples_per_second": 12.628,
      "eval_steps_per_second": 1.579,
      "step": 80000
    },
    {
      "epoch": 2.361111111111111,
      "grad_norm": 0.9593023061752319,
      "learning_rate": 7.5e-05,
      "loss": 1.408,
      "step": 85000
    },
    {
      "epoch": 2.5,
      "grad_norm": 1.440004825592041,
      "learning_rate": 5e-05,
      "loss": 1.3956,
      "step": 90000
    },
    {
      "epoch": 2.5,
      "eval_bleu": 23.3848,
      "eval_gen_len": 27.6637,
      "eval_loss": 1.752461552619934,
      "eval_runtime": 237.0184,
      "eval_samples_per_second": 12.657,
      "eval_steps_per_second": 1.582,
      "step": 90000
    },
    {
      "epoch": 2.638888888888889,
      "grad_norm": 1.1929932832717896,
      "learning_rate": 2.5e-05,
      "loss": 1.3938,
      "step": 95000
    },
    {
      "epoch": 2.7777777777777777,
      "grad_norm": 1.0216492414474487,
      "learning_rate": 0.0,
      "loss": 1.3903,
      "step": 100000
    },
    {
      "epoch": 2.7777777777777777,
      "eval_bleu": 23.6596,
      "eval_gen_len": 27.526,
      "eval_loss": 1.7469114065170288,
      "eval_runtime": 235.9542,
      "eval_samples_per_second": 12.714,
      "eval_steps_per_second": 1.589,
      "step": 100000
    },
    {
      "epoch": 2.7777777777777777,
      "step": 100000,
      "total_flos": 3.803274433029734e+16,
      "train_loss": 1.5316169482421875,
      "train_runtime": 15895.0874,
      "train_samples_per_second": 100.66,
      "train_steps_per_second": 6.291
    }
  ],
  "logging_steps": 5000,
  "max_steps": 100000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 10000,
  "total_flos": 3.803274433029734e+16,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}