trainer_state.json

{
  "best_metric": 0.4794759750366211,
  "best_model_checkpoint": "hubert-base-tamil-ipa/checkpoint-1500",
  "epoch": 1.0706512042818912,
  "global_step": 1500,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.07,
      "learning_rate": 2.91e-05,
      "loss": 5.9666,
      "step": 100
    },
    {
      "epoch": 0.14,
      "learning_rate": 5.88e-05,
      "loss": 2.691,
      "step": 200
    },
    {
      "epoch": 0.21,
      "learning_rate": 8.879999999999999e-05,
      "loss": 2.5892,
      "step": 300
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0001185,
      "loss": 2.276,
      "step": 400
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0001482,
      "loss": 0.9141,
      "step": 500
    },
    {
      "epoch": 0.36,
      "eval_cer": 0.41183358071245185,
      "eval_loss": 0.7568618655204773,
      "eval_runtime": 1088.5554,
      "eval_samples_per_second": 11.069,
      "eval_steps_per_second": 1.384,
      "step": 500
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0001779,
      "loss": 0.5126,
      "step": 600
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00020759999999999998,
      "loss": 0.431,
      "step": 700
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002376,
      "loss": 0.3866,
      "step": 800
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.000267,
      "loss": 0.3897,
      "step": 900
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00029699999999999996,
      "loss": 0.3504,
      "step": 1000
    },
    {
      "epoch": 0.71,
      "eval_cer": 0.3094838133825417,
      "eval_loss": 0.5331771373748779,
      "eval_runtime": 1206.7225,
      "eval_samples_per_second": 9.985,
      "eval_steps_per_second": 1.249,
      "step": 1000
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00029900074019245003,
      "loss": 0.3487,
      "step": 1100
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002979126572908956,
      "loss": 0.3302,
      "step": 1200
    },
    {
      "epoch": 0.93,
      "learning_rate": 0.0002968245743893412,
      "loss": 0.3246,
      "step": 1300
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0002957142857142857,
      "loss": 0.301,
      "step": 1400
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00029460399703923015,
      "loss": 0.2829,
      "step": 1500
    },
    {
      "epoch": 1.07,
      "eval_cer": 0.2748139642169448,
      "eval_loss": 0.4794759750366211,
      "eval_runtime": 1193.809,
      "eval_samples_per_second": 10.093,
      "eval_steps_per_second": 1.262,
      "step": 1500
    }
  ],
  "max_steps": 28020,
  "num_train_epochs": 20,
  "total_flos": 2.852051261361667e+18,
  "trial_name": null,
  "trial_params": null
}