trainer_state.json

{
  "best_global_step": 1500,
  "best_metric": 0.33901429176330566,
  "best_model_checkpoint": "./stt/checkpoint-1500",
  "epoch": 1.854140914709518,
  "eval_steps": 500,
  "global_step": 1500,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.12360939431396786,
      "grad_norm": 10.66793155670166,
      "learning_rate": 9.999173944549022e-06,
      "loss": 1.6423,
      "step": 100
    },
    {
      "epoch": 0.24721878862793573,
      "grad_norm": 8.940858840942383,
      "learning_rate": 9.991680624320292e-06,
      "loss": 0.5818,
      "step": 200
    },
    {
      "epoch": 0.37082818294190356,
      "grad_norm": 7.630367755889893,
      "learning_rate": 9.976393387344834e-06,
      "loss": 0.4997,
      "step": 300
    },
    {
      "epoch": 0.49443757725587145,
      "grad_norm": 8.54523754119873,
      "learning_rate": 9.953336102828482e-06,
      "loss": 0.4622,
      "step": 400
    },
    {
      "epoch": 0.6180469715698393,
      "grad_norm": 8.749971389770508,
      "learning_rate": 9.922544771983945e-06,
      "loss": 0.4135,
      "step": 500
    },
    {
      "epoch": 0.6180469715698393,
      "eval_loss": 0.4068942070007324,
      "eval_runtime": 1305.4198,
      "eval_samples_per_second": 2.477,
      "eval_steps_per_second": 0.156,
      "eval_wer": 21.631878557874764,
      "eval_wer_ortho": 29.911516425399192,
      "step": 500
    },
    {
      "epoch": 0.7416563658838071,
      "grad_norm": 7.349922180175781,
      "learning_rate": 9.884067471819184e-06,
      "loss": 0.3882,
      "step": 600
    },
    {
      "epoch": 0.865265760197775,
      "grad_norm": 8.007939338684082,
      "learning_rate": 9.837964280070878e-06,
      "loss": 0.3613,
      "step": 700
    },
    {
      "epoch": 0.9888751545117429,
      "grad_norm": 7.204431056976318,
      "learning_rate": 9.784307181400159e-06,
      "loss": 0.3561,
      "step": 800
    },
    {
      "epoch": 1.1124845488257107,
      "grad_norm": 6.045726776123047,
      "learning_rate": 9.723179954997127e-06,
      "loss": 0.2156,
      "step": 900
    },
    {
      "epoch": 1.2360939431396787,
      "grad_norm": 5.328393459320068,
      "learning_rate": 9.654678043769566e-06,
      "loss": 0.2036,
      "step": 1000
    },
    {
      "epoch": 1.2360939431396787,
      "eval_loss": 0.35839423537254333,
      "eval_runtime": 1301.6256,
      "eval_samples_per_second": 2.484,
      "eval_steps_per_second": 0.156,
      "eval_wer": 18.355245323936025,
      "eval_wer_ortho": 25.873817676373868,
      "step": 1000
    },
    {
      "epoch": 1.3597033374536465,
      "grad_norm": 6.864717960357666,
      "learning_rate": 9.578908405320196e-06,
      "loss": 0.2019,
      "step": 1100
    },
    {
      "epoch": 1.4833127317676142,
      "grad_norm": 5.546298027038574,
      "learning_rate": 9.495989344945056e-06,
      "loss": 0.211,
      "step": 1200
    },
    {
      "epoch": 1.6069221260815822,
      "grad_norm": 4.569760322570801,
      "learning_rate": 9.406050330913839e-06,
      "loss": 0.196,
      "step": 1300
    },
    {
      "epoch": 1.73053152039555,
      "grad_norm": 6.44041109085083,
      "learning_rate": 9.309231792320548e-06,
      "loss": 0.2003,
      "step": 1400
    },
    {
      "epoch": 1.854140914709518,
      "grad_norm": 6.337347507476807,
      "learning_rate": 9.205684899820146e-06,
      "loss": 0.1899,
      "step": 1500
    },
    {
      "epoch": 1.854140914709518,
      "eval_loss": 0.33901429176330566,
      "eval_runtime": 1299.601,
      "eval_samples_per_second": 2.488,
      "eval_steps_per_second": 0.156,
      "eval_wer": 16.481431282190297,
      "eval_wer_ortho": 24.09397565854155,
      "step": 1500
    }
  ],
  "logging_steps": 100,
  "max_steps": 8000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 10,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 2.448022931177472e+19,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}