{
  "best_metric": 42.972,
  "best_model_checkpoint": "opus_big_wce_random/checkpoint-40000",
  "epoch": 0.9858287122612446,
  "global_step": 40000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.2,
      "learning_rate": 1.9753635243376465e-05,
      "loss": 0.181,
      "step": 8000
    },
    {
      "epoch": 0.2,
      "eval_bleu": 42.0536,
      "eval_gen_len": 33.651,
      "eval_loss": 0.09812351316213608,
      "eval_runtime": 117.8669,
      "eval_samples_per_second": 8.849,
      "eval_steps_per_second": 0.56,
      "step": 8000
    },
    {
      "epoch": 0.39,
      "learning_rate": 1.9507208872458414e-05,
      "loss": 0.1503,
      "step": 16000
    },
    {
      "epoch": 0.39,
      "eval_bleu": 42.8859,
      "eval_gen_len": 34.6251,
      "eval_loss": 0.09743236750364304,
      "eval_runtime": 145.7149,
      "eval_samples_per_second": 7.158,
      "eval_steps_per_second": 0.453,
      "step": 16000
    },
    {
      "epoch": 0.59,
      "learning_rate": 1.9260844115834877e-05,
      "loss": 0.1461,
      "step": 24000
    },
    {
      "epoch": 0.59,
      "eval_bleu": 42.882,
      "eval_gen_len": 33.5225,
      "eval_loss": 0.09730042517185211,
      "eval_runtime": 187.569,
      "eval_samples_per_second": 5.561,
      "eval_steps_per_second": 0.352,
      "step": 24000
    },
    {
      "epoch": 0.79,
      "learning_rate": 1.901447935921134e-05,
      "loss": 0.1422,
      "step": 32000
    },
    {
      "epoch": 0.79,
      "eval_bleu": 42.9206,
      "eval_gen_len": 34.2148,
      "eval_loss": 0.09637172520160675,
      "eval_runtime": 116.8881,
      "eval_samples_per_second": 8.923,
      "eval_steps_per_second": 0.565,
      "step": 32000
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.8768114602587803e-05,
      "loss": 0.1407,
      "step": 40000
    },
    {
      "epoch": 0.99,
      "eval_bleu": 42.972,
      "eval_gen_len": 34.4257,
      "eval_loss": 0.09560249000787735,
      "eval_runtime": 125.6435,
      "eval_samples_per_second": 8.301,
      "eval_steps_per_second": 0.525,
      "step": 40000
    }
  ],
  "max_steps": 649200,
  "num_train_epochs": 16,
  "total_flos": 1.2920984272738714e+17,
  "trial_name": null,
  "trial_params": null
}