{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 15.0,
  "global_step": 20085,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "learning_rate": 9.828625334101986e-05,
      "loss": 11.016,
      "step": 1339
    },
    {
      "epoch": 1.0,
      "eval_cer": 0.24761904761904763,
      "eval_loss": 1.0964484214782715,
      "eval_new_wer": 0.12380952380952381,
      "eval_old_wer": 0.7619047619047619,
      "eval_runtime": 7.9497,
      "eval_samples_per_second": 26.416,
      "eval_steps_per_second": 3.396,
      "step": 1339
    },
    {
      "epoch": 2.0,
      "learning_rate": 9.126880142550182e-05,
      "loss": 0.5219,
      "step": 2678
    },
    {
      "epoch": 2.0,
      "eval_cer": 0.15844155844155844,
      "eval_loss": 1.1968157291412354,
      "eval_new_wer": 0.13333333333333333,
      "eval_old_wer": 0.3619047619047619,
      "eval_runtime": 7.7729,
      "eval_samples_per_second": 27.017,
      "eval_steps_per_second": 3.474,
      "step": 2678
    },
    {
      "epoch": 3.0,
      "learning_rate": 8.425134950998376e-05,
      "loss": 0.2575,
      "step": 4017
    },
    {
      "epoch": 3.0,
      "eval_cer": 0.18095238095238095,
      "eval_loss": 1.9693397283554077,
      "eval_new_wer": 0.18095238095238095,
      "eval_old_wer": 0.41904761904761906,
      "eval_runtime": 7.7377,
      "eval_samples_per_second": 27.14,
      "eval_steps_per_second": 3.489,
      "step": 4017
    },
    {
      "epoch": 4.0,
      "learning_rate": 7.72338975944657e-05,
      "loss": 0.2088,
      "step": 5356
    },
    {
      "epoch": 4.0,
      "eval_cer": 0.14112554112554113,
      "eval_loss": 1.7389041185379028,
      "eval_new_wer": 0.1761904761904762,
      "eval_old_wer": 0.2619047619047619,
      "eval_runtime": 7.7859,
      "eval_samples_per_second": 26.972,
      "eval_steps_per_second": 3.468,
      "step": 5356
    },
    {
      "epoch": 5.0,
      "learning_rate": 7.021644567894765e-05,
      "loss": 0.1735,
      "step": 6695
    },
    {
      "epoch": 5.0,
      "eval_cer": 0.14632034632034632,
      "eval_loss": 1.926347017288208,
      "eval_new_wer": 0.11428571428571428,
      "eval_old_wer": 0.2714285714285714,
      "eval_runtime": 7.8378,
      "eval_samples_per_second": 26.793,
      "eval_steps_per_second": 3.445,
      "step": 6695
    },
    {
      "epoch": 6.0,
      "learning_rate": 6.319899376342959e-05,
      "loss": 0.1382,
      "step": 8034
    },
    {
      "epoch": 6.0,
      "eval_cer": 0.129004329004329,
      "eval_loss": 1.5699124336242676,
      "eval_new_wer": 0.13333333333333333,
      "eval_old_wer": 0.23809523809523808,
      "eval_runtime": 7.8478,
      "eval_samples_per_second": 26.759,
      "eval_steps_per_second": 3.44,
      "step": 8034
    },
    {
      "epoch": 7.0,
      "learning_rate": 5.6181541847911533e-05,
      "loss": 0.1213,
      "step": 9373
    },
    {
      "epoch": 7.0,
      "eval_cer": 0.12207792207792208,
      "eval_loss": 1.3351221084594727,
      "eval_new_wer": 0.14285714285714285,
      "eval_old_wer": 0.23333333333333334,
      "eval_runtime": 7.7317,
      "eval_samples_per_second": 27.161,
      "eval_steps_per_second": 3.492,
      "step": 9373
    },
    {
      "epoch": 8.0,
      "learning_rate": 4.916408993239348e-05,
      "loss": 0.0967,
      "step": 10712
    },
    {
      "epoch": 8.0,
      "eval_cer": 0.11255411255411256,
      "eval_loss": 1.5987052917480469,
      "eval_new_wer": 0.11904761904761904,
      "eval_old_wer": 0.23809523809523808,
      "eval_runtime": 7.7718,
      "eval_samples_per_second": 27.021,
      "eval_steps_per_second": 3.474,
      "step": 10712
    },
    {
      "epoch": 9.0,
      "learning_rate": 4.2151878832346316e-05,
      "loss": 0.0883,
      "step": 12051
    },
    {
      "epoch": 9.0,
      "eval_cer": 0.1774891774891775,
      "eval_loss": 3.794250011444092,
      "eval_new_wer": 0.23333333333333334,
      "eval_old_wer": 0.3047619047619048,
      "eval_runtime": 7.8447,
      "eval_samples_per_second": 26.77,
      "eval_steps_per_second": 3.442,
      "step": 12051
    },
    {
      "epoch": 10.0,
      "learning_rate": 3.5134426916828264e-05,
      "loss": 0.0711,
      "step": 13390
    },
    {
      "epoch": 10.0,
      "eval_cer": 0.11255411255411256,
      "eval_loss": 1.1626664400100708,
      "eval_new_wer": 0.10476190476190476,
      "eval_old_wer": 0.23333333333333334,
      "eval_runtime": 7.7431,
      "eval_samples_per_second": 27.121,
      "eval_steps_per_second": 3.487,
      "step": 13390
    },
    {
      "epoch": 11.0,
      "learning_rate": 2.8116975001310204e-05,
      "loss": 0.0571,
      "step": 14729
    },
    {
      "epoch": 11.0,
      "eval_cer": 0.11688311688311688,
      "eval_loss": 0.9811009168624878,
      "eval_new_wer": 0.10952380952380952,
      "eval_old_wer": 0.21904761904761905,
      "eval_runtime": 7.8259,
      "eval_samples_per_second": 26.834,
      "eval_steps_per_second": 3.45,
      "step": 14729
    },
    {
      "epoch": 12.0,
      "learning_rate": 2.110476390126304e-05,
      "loss": 0.0482,
      "step": 16068
    },
    {
      "epoch": 12.0,
      "eval_cer": 0.11428571428571428,
      "eval_loss": 1.093704104423523,
      "eval_new_wer": 0.10476190476190476,
      "eval_old_wer": 0.22857142857142856,
      "eval_runtime": 7.8884,
      "eval_samples_per_second": 26.621,
      "eval_steps_per_second": 3.423,
      "step": 16068
    },
    {
      "epoch": 13.0,
      "learning_rate": 1.409255280121587e-05,
      "loss": 0.0379,
      "step": 17407
    },
    {
      "epoch": 13.0,
      "eval_cer": 0.14891774891774892,
      "eval_loss": 1.6621599197387695,
      "eval_new_wer": 0.11428571428571428,
      "eval_old_wer": 0.26666666666666666,
      "eval_runtime": 7.746,
      "eval_samples_per_second": 27.111,
      "eval_steps_per_second": 3.486,
      "step": 17407
    },
    {
      "epoch": 14.0,
      "learning_rate": 7.0751008856978145e-06,
      "loss": 0.0385,
      "step": 18746
    },
    {
      "epoch": 14.0,
      "eval_cer": 0.129004329004329,
      "eval_loss": 1.4713504314422607,
      "eval_new_wer": 0.12857142857142856,
      "eval_old_wer": 0.2523809523809524,
      "eval_runtime": 7.8845,
      "eval_samples_per_second": 26.635,
      "eval_steps_per_second": 3.424,
      "step": 18746
    },
    {
      "epoch": 15.0,
      "learning_rate": 5.764897017975998e-08,
      "loss": 0.0255,
      "step": 20085
    },
    {
      "epoch": 15.0,
      "eval_cer": 0.12987012987012986,
      "eval_loss": 1.6889647245407104,
      "eval_new_wer": 0.13333333333333333,
      "eval_old_wer": 0.24285714285714285,
      "eval_runtime": 7.9672,
      "eval_samples_per_second": 26.358,
      "eval_steps_per_second": 3.389,
      "step": 20085
    }
  ],
  "max_steps": 20085,
  "num_train_epochs": 15,
  "total_flos": 1.5248746564040909e+19,
  "trial_name": null,
  "trial_params": null
}