trainer_state.json

{
  "best_metric": 0.3982403874397278,
  "best_model_checkpoint": "XLS-R-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": 2.94e-05,
      "loss": 7.2232,
      "step": 100
    },
    {
      "epoch": 0.14,
      "learning_rate": 5.91e-05,
      "loss": 3.0769,
      "step": 200
    },
    {
      "epoch": 0.21,
      "learning_rate": 8.909999999999998e-05,
      "loss": 2.616,
      "step": 300
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0001188,
      "loss": 2.3839,
      "step": 400
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00014849999999999998,
      "loss": 1.0712,
      "step": 500
    },
    {
      "epoch": 0.36,
      "eval_cer": 0.3590590901797176,
      "eval_loss": 0.6315754652023315,
      "eval_runtime": 1320.9999,
      "eval_samples_per_second": 9.121,
      "eval_steps_per_second": 1.141,
      "step": 500
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00017819999999999997,
      "loss": 0.4126,
      "step": 600
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00020789999999999996,
      "loss": 0.3474,
      "step": 700
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00023789999999999998,
      "loss": 0.3079,
      "step": 800
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002673,
      "loss": 0.3011,
      "step": 900
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00029729999999999996,
      "loss": 0.2795,
      "step": 1000
    },
    {
      "epoch": 0.71,
      "eval_cer": 0.25890564826700896,
      "eval_loss": 0.44443580508232117,
      "eval_runtime": 1282.5832,
      "eval_samples_per_second": 9.394,
      "eval_steps_per_second": 1.175,
      "step": 1000
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00029898963730569947,
      "loss": 0.2838,
      "step": 1100
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00029790155440414503,
      "loss": 0.2674,
      "step": 1200
    },
    {
      "epoch": 0.93,
      "learning_rate": 0.00029681347150259065,
      "loss": 0.2531,
      "step": 1300
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00029570318282753515,
      "loss": 0.2508,
      "step": 1400
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0002945928941524796,
      "loss": 0.2403,
      "step": 1500
    },
    {
      "epoch": 1.07,
      "eval_cer": 0.2333796333440308,
      "eval_loss": 0.3982403874397278,
      "eval_runtime": 1264.6744,
      "eval_samples_per_second": 9.527,
      "eval_steps_per_second": 1.192,
      "step": 1500
    }
  ],
  "max_steps": 28020,
  "num_train_epochs": 20,
  "total_flos": 9.531198923471569e+18,
  "trial_name": null,
  "trial_params": null
}