trainer_state.json

{
  "best_metric": 0.34306156635284424,
  "best_model_checkpoint": "XLS-R-tamil-phoneme-v2/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": 2.97e-05,
      "loss": 7.2256,
      "step": 100
    },
    {
      "epoch": 0.14,
      "learning_rate": 5.94e-05,
      "loss": 3.0456,
      "step": 200
    },
    {
      "epoch": 0.21,
      "learning_rate": 8.939999999999999e-05,
      "loss": 2.6409,
      "step": 300
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0001191,
      "loss": 2.4601,
      "step": 400
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00014879999999999998,
      "loss": 1.3437,
      "step": 500
    },
    {
      "epoch": 0.36,
      "eval_cer": 0.402669688703466,
      "eval_loss": 0.73838210105896,
      "eval_runtime": 807.8476,
      "eval_samples_per_second": 14.915,
      "eval_steps_per_second": 1.865,
      "step": 500
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00017849999999999997,
      "loss": 0.449,
      "step": 600
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00020819999999999996,
      "loss": 0.3495,
      "step": 700
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002382,
      "loss": 0.3063,
      "step": 800
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002676,
      "loss": 0.3021,
      "step": 900
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00029759999999999997,
      "loss": 0.2685,
      "step": 1000
    },
    {
      "epoch": 0.71,
      "eval_cer": 0.2228818998716303,
      "eval_loss": 0.3911130428314209,
      "eval_runtime": 1044.7679,
      "eval_samples_per_second": 11.533,
      "eval_steps_per_second": 1.442,
      "step": 1000
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00029980158159597406,
      "loss": 0.2701,
      "step": 1100
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00029959022286125085,
      "loss": 0.2527,
      "step": 1200
    },
    {
      "epoch": 0.93,
      "learning_rate": 0.00029937886412652765,
      "loss": 0.2472,
      "step": 1300
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00029916319194823863,
      "loss": 0.2317,
      "step": 1400
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.00029894751976994966,
      "loss": 0.2197,
      "step": 1500
    },
    {
      "epoch": 1.07,
      "eval_cer": 0.20153040757381258,
      "eval_loss": 0.34306156635284424,
      "eval_runtime": 1175.3837,
      "eval_samples_per_second": 10.251,
      "eval_steps_per_second": 1.282,
      "step": 1500
    }
  ],
  "max_steps": 140100,
  "num_train_epochs": 100,
  "total_flos": 9.531198923471569e+18,
  "trial_name": null,
  "trial_params": null
}