trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 2.999398677089597,
  "eval_steps": 500,
  "global_step": 7482,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.20044097013429546,
      "grad_norm": 6.546669006347656,
      "learning_rate": 0.0001335113484646195,
      "loss": 4.9792,
      "step": 500
    },
    {
      "epoch": 0.4008819402685909,
      "grad_norm": 12.99987506866455,
      "learning_rate": 0.00019254418535571068,
      "loss": 0.9756,
      "step": 1000
    },
    {
      "epoch": 0.6013229104028863,
      "grad_norm": 5.031010627746582,
      "learning_rate": 0.00017769196494875984,
      "loss": 0.7285,
      "step": 1500
    },
    {
      "epoch": 0.8017638805371818,
      "grad_norm": 5.174665451049805,
      "learning_rate": 0.000162839744541809,
      "loss": 0.6145,
      "step": 2000
    },
    {
      "epoch": 0.9997995590298657,
      "eval_cer": 0.06818140975752038,
      "eval_loss": 0.2314606010913849,
      "eval_runtime": 598.1802,
      "eval_samples_per_second": 59.596,
      "eval_steps_per_second": 7.451,
      "eval_wer": 0.08141925042668663,
      "step": 2494
    },
    {
      "epoch": 1.0022048506714774,
      "grad_norm": 1.5678229331970215,
      "learning_rate": 0.00014798752413485816,
      "loss": 0.5526,
      "step": 2500
    },
    {
      "epoch": 1.2026458208057726,
      "grad_norm": 1.3437103033065796,
      "learning_rate": 0.00013313530372790732,
      "loss": 0.4812,
      "step": 3000
    },
    {
      "epoch": 1.4030867909400682,
      "grad_norm": 1.1184632778167725,
      "learning_rate": 0.00011828308332095648,
      "loss": 0.4596,
      "step": 3500
    },
    {
      "epoch": 1.6035277610743637,
      "grad_norm": 1.1852930784225464,
      "learning_rate": 0.00010343086291400565,
      "loss": 0.4337,
      "step": 4000
    },
    {
      "epoch": 1.8039687312086592,
      "grad_norm": 1.735134243965149,
      "learning_rate": 8.857864250705481e-05,
      "loss": 0.4098,
      "step": 4500
    },
    {
      "epoch": 2.0,
      "eval_cer": 0.056794456555446654,
      "eval_loss": 0.17900477349758148,
      "eval_runtime": 419.809,
      "eval_samples_per_second": 84.917,
      "eval_steps_per_second": 10.617,
      "eval_wer": 0.06738346920647913,
      "step": 4989
    },
    {
      "epoch": 2.0044097013429547,
      "grad_norm": 0.9114906191825867,
      "learning_rate": 7.372642210010397e-05,
      "loss": 0.3887,
      "step": 5000
    },
    {
      "epoch": 2.20485067147725,
      "grad_norm": 1.5065760612487793,
      "learning_rate": 5.887420169315313e-05,
      "loss": 0.3298,
      "step": 5500
    },
    {
      "epoch": 2.4052916416115453,
      "grad_norm": 1.4358354806900024,
      "learning_rate": 4.402198128620229e-05,
      "loss": 0.3172,
      "step": 6000
    },
    {
      "epoch": 2.605732611745841,
      "grad_norm": 0.6124329566955566,
      "learning_rate": 2.9169760879251452e-05,
      "loss": 0.3094,
      "step": 6500
    },
    {
      "epoch": 2.8061735818801363,
      "grad_norm": 1.8411574363708496,
      "learning_rate": 1.4317540472300609e-05,
      "loss": 0.2986,
      "step": 7000
    },
    {
      "epoch": 2.999398677089597,
      "eval_cer": 0.04912028966875449,
      "eval_loss": 0.14919918775558472,
      "eval_runtime": 448.6832,
      "eval_samples_per_second": 79.453,
      "eval_steps_per_second": 9.934,
      "eval_wer": 0.0574731780364509,
      "step": 7482
    },
    {
      "epoch": 2.999398677089597,
      "step": 7482,
      "total_flos": 4.06242496282012e+19,
      "train_loss": 0.7722781848474098,
      "train_runtime": 4057.7267,
      "train_samples_per_second": 59.015,
      "train_steps_per_second": 1.844
    }
  ],
  "logging_steps": 500,
  "max_steps": 7482,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 4.06242496282012e+19,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}