{
  "best_metric": 43.8722,
  "best_model_checkpoint": "opus_big_adapt_wce_gloss_train-sampled_bands_5_ubweight_1.75/checkpoint-80000",
  "epoch": 1.9716574245224892,
  "global_step": 80000,
  "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": 0.1776,
      "step": 8000
    },
    {
      "epoch": 0.2,
      "eval_bleu": 42.0642,
      "eval_gen_len": 33.7881,
      "eval_loss": 0.09869997948408127,
      "eval_runtime": 84.9196,
      "eval_samples_per_second": 12.282,
      "eval_steps_per_second": 0.777,
      "step": 8000
    },
    {
      "epoch": 0.39,
      "learning_rate": 1.950723967960567e-05,
      "loss": 0.1467,
      "step": 16000
    },
    {
      "epoch": 0.39,
      "eval_bleu": 41.7526,
      "eval_gen_len": 34.7795,
      "eval_loss": 0.09779118001461029,
      "eval_runtime": 93.0132,
      "eval_samples_per_second": 11.213,
      "eval_steps_per_second": 0.71,
      "step": 16000
    },
    {
      "epoch": 0.59,
      "learning_rate": 1.926090573012939e-05,
      "loss": 0.1426,
      "step": 24000
    },
    {
      "epoch": 0.59,
      "eval_bleu": 42.0584,
      "eval_gen_len": 33.464,
      "eval_loss": 0.09806442260742188,
      "eval_runtime": 80.6234,
      "eval_samples_per_second": 12.937,
      "eval_steps_per_second": 0.819,
      "step": 24000
    },
    {
      "epoch": 0.79,
      "learning_rate": 1.9014540973505854e-05,
      "loss": 0.1387,
      "step": 32000
    },
    {
      "epoch": 0.79,
      "eval_bleu": 42.6266,
      "eval_gen_len": 33.5254,
      "eval_loss": 0.09653611481189728,
      "eval_runtime": 81.9234,
      "eval_samples_per_second": 12.731,
      "eval_steps_per_second": 0.806,
      "step": 32000
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.876814540973506e-05,
      "loss": 0.1373,
      "step": 40000
    },
    {
      "epoch": 0.99,
      "eval_bleu": 42.6264,
      "eval_gen_len": 33.9195,
      "eval_loss": 0.09597118198871613,
      "eval_runtime": 81.6116,
      "eval_samples_per_second": 12.78,
      "eval_steps_per_second": 0.809,
      "step": 40000
    },
    {
      "epoch": 1.18,
      "learning_rate": 1.852178065311152e-05,
      "loss": 0.1222,
      "step": 48000
    },
    {
      "epoch": 1.18,
      "eval_bleu": 42.7749,
      "eval_gen_len": 34.1208,
      "eval_loss": 0.09734106063842773,
      "eval_runtime": 94.7556,
      "eval_samples_per_second": 11.007,
      "eval_steps_per_second": 0.697,
      "step": 48000
    },
    {
      "epoch": 1.38,
      "learning_rate": 1.8275415896487988e-05,
      "loss": 0.1226,
      "step": 56000
    },
    {
      "epoch": 1.38,
      "eval_bleu": 42.687,
      "eval_gen_len": 34.1697,
      "eval_loss": 0.09724980592727661,
      "eval_runtime": 191.9924,
      "eval_samples_per_second": 5.433,
      "eval_steps_per_second": 0.344,
      "step": 56000
    },
    {
      "epoch": 1.58,
      "learning_rate": 1.802905113986445e-05,
      "loss": 0.1233,
      "step": 64000
    },
    {
      "epoch": 1.58,
      "eval_bleu": 42.9609,
      "eval_gen_len": 34.3432,
      "eval_loss": 0.09664924442768097,
      "eval_runtime": 125.3853,
      "eval_samples_per_second": 8.318,
      "eval_steps_per_second": 0.526,
      "step": 64000
    },
    {
      "epoch": 1.77,
      "learning_rate": 1.7782717190388173e-05,
      "loss": 0.1227,
      "step": 72000
    },
    {
      "epoch": 1.77,
      "eval_bleu": 42.7711,
      "eval_gen_len": 34.1985,
      "eval_loss": 0.09645862132310867,
      "eval_runtime": 85.6207,
      "eval_samples_per_second": 12.182,
      "eval_steps_per_second": 0.771,
      "step": 72000
    },
    {
      "epoch": 1.97,
      "learning_rate": 1.7536321626617377e-05,
      "loss": 0.1224,
      "step": 80000
    },
    {
      "epoch": 1.97,
      "eval_bleu": 43.8722,
      "eval_gen_len": 33.5513,
      "eval_loss": 0.09601087123155594,
      "eval_runtime": 85.5385,
      "eval_samples_per_second": 12.193,
      "eval_steps_per_second": 0.772,
      "step": 80000
    }
  ],
  "max_steps": 649200,
  "num_train_epochs": 16,
  "total_flos": 2.5844916958632346e+17,
  "trial_name": null,
  "trial_params": null
}