checkpoint-2328/trainer_state.json

{
  "best_global_step": 1200,
  "best_metric": 8.762886597938143,
  "best_model_checkpoint": "./ZINASR/checkpoint-1200",
  "epoch": 8.0,
  "eval_steps": 400,
  "global_step": 2328,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.6872852233676976,
      "grad_norm": 5.297506809234619,
      "learning_rate": 9.950000000000001e-06,
      "loss": 0.1217,
      "step": 200
    },
    {
      "epoch": 1.3745704467353952,
      "grad_norm": 1.6356655359268188,
      "learning_rate": 9.06484962406015e-06,
      "loss": 0.0826,
      "step": 400
    },
    {
      "epoch": 1.3745704467353952,
      "eval_loss": 0.1176551878452301,
      "eval_runtime": 1451.5411,
      "eval_samples_per_second": 1.602,
      "eval_steps_per_second": 0.2,
      "eval_wer": 8.891752577319588,
      "step": 400
    },
    {
      "epoch": 2.0618556701030926,
      "grad_norm": 0.9965283274650574,
      "learning_rate": 8.134398496240602e-06,
      "loss": 0.0483,
      "step": 600
    },
    {
      "epoch": 2.7491408934707904,
      "grad_norm": 1.6627748012542725,
      "learning_rate": 7.194548872180452e-06,
      "loss": 0.018,
      "step": 800
    },
    {
      "epoch": 2.7491408934707904,
      "eval_loss": 0.121592216193676,
      "eval_runtime": 1448.1115,
      "eval_samples_per_second": 1.606,
      "eval_steps_per_second": 0.201,
      "eval_wer": 9.211531118747613,
      "step": 800
    },
    {
      "epoch": 3.436426116838488,
      "grad_norm": 1.0129088163375854,
      "learning_rate": 6.254699248120302e-06,
      "loss": 0.0118,
      "step": 1000
    },
    {
      "epoch": 4.123711340206185,
      "grad_norm": 0.18754775822162628,
      "learning_rate": 5.314849624060151e-06,
      "loss": 0.0067,
      "step": 1200
    },
    {
      "epoch": 4.123711340206185,
      "eval_loss": 0.12416346371173859,
      "eval_runtime": 1449.1081,
      "eval_samples_per_second": 1.604,
      "eval_steps_per_second": 0.201,
      "eval_wer": 8.762886597938143,
      "step": 1200
    },
    {
      "epoch": 4.810996563573883,
      "grad_norm": 0.20880936086177826,
      "learning_rate": 4.3750000000000005e-06,
      "loss": 0.0034,
      "step": 1400
    },
    {
      "epoch": 5.498281786941581,
      "grad_norm": 0.06731250137090683,
      "learning_rate": 3.43515037593985e-06,
      "loss": 0.0022,
      "step": 1600
    },
    {
      "epoch": 5.498281786941581,
      "eval_loss": 0.12590037286281586,
      "eval_runtime": 1450.6425,
      "eval_samples_per_second": 1.603,
      "eval_steps_per_second": 0.201,
      "eval_wer": 8.858342878961436,
      "step": 1600
    },
    {
      "epoch": 6.185567010309279,
      "grad_norm": 0.056367386132478714,
      "learning_rate": 2.4953007518796995e-06,
      "loss": 0.0014,
      "step": 1800
    },
    {
      "epoch": 6.872852233676976,
      "grad_norm": 0.0637979805469513,
      "learning_rate": 1.5554511278195489e-06,
      "loss": 0.0012,
      "step": 2000
    },
    {
      "epoch": 6.872852233676976,
      "eval_loss": 0.12869131565093994,
      "eval_runtime": 1446.4517,
      "eval_samples_per_second": 1.607,
      "eval_steps_per_second": 0.201,
      "eval_wer": 8.858342878961436,
      "step": 2000
    },
    {
      "epoch": 7.560137457044673,
      "grad_norm": 0.0420345701277256,
      "learning_rate": 6.156015037593985e-07,
      "loss": 0.001,
      "step": 2200
    }
  ],
  "logging_steps": 200,
  "max_steps": 2328,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 8,
  "save_steps": 1200,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 2.146844510060544e+19,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}