trainer_state.json

{
  "best_global_step": 5570,
  "best_metric": 0.7717538371980056,
  "best_model_checkpoint": "./checkpoints/checkpoint-5570",
  "epoch": 0.9998653923812088,
  "eval_steps": 1114,
  "global_step": 5571,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0897384125274824,
      "grad_norm": 0.10910310596227646,
      "learning_rate": 4.901672070547205e-05,
      "loss": 2.0607,
      "step": 500
    },
    {
      "epoch": 0.1794768250549648,
      "grad_norm": 0.11283092200756073,
      "learning_rate": 4.6136704554635066e-05,
      "loss": 0.1219,
      "step": 1000
    },
    {
      "epoch": 0.19993718311123077,
      "eval_cer": 0.7744026050691318,
      "eval_loss": 0.06905069202184677,
      "eval_runtime": 2200.6581,
      "eval_samples_per_second": 80.463,
      "eval_steps_per_second": 0.21,
      "step": 1114
    },
    {
      "epoch": 0.26921523758244714,
      "grad_norm": 0.07702942937612534,
      "learning_rate": 4.9413107833705604e-05,
      "loss": 0.0709,
      "step": 1500
    },
    {
      "epoch": 0.3589536501099296,
      "grad_norm": 0.048379454761743546,
      "learning_rate": 4.695072154439477e-05,
      "loss": 0.0526,
      "step": 2000
    },
    {
      "epoch": 0.39987436622246153,
      "eval_cer": 0.7726060406484112,
      "eval_loss": 0.03332927078008652,
      "eval_runtime": 436.8886,
      "eval_samples_per_second": 405.3,
      "eval_steps_per_second": 1.057,
      "step": 2228
    },
    {
      "epoch": 0.44869206263741196,
      "grad_norm": 0.04356146231293678,
      "learning_rate": 4.970863550514242e-05,
      "loss": 0.0398,
      "step": 2500
    },
    {
      "epoch": 0.5384304751648943,
      "grad_norm": 0.03718848526477814,
      "learning_rate": 4.767405209364289e-05,
      "loss": 0.0348,
      "step": 3000
    },
    {
      "epoch": 0.5998115493336923,
      "eval_cer": 0.7720533296293449,
      "eval_loss": 0.022723019123077393,
      "eval_runtime": 439.0491,
      "eval_samples_per_second": 403.306,
      "eval_steps_per_second": 1.052,
      "step": 3342
    },
    {
      "epoch": 0.6281688876923768,
      "grad_norm": 0.03279432654380798,
      "learning_rate": 4.384876624004233e-05,
      "loss": 0.03,
      "step": 3500
    },
    {
      "epoch": 0.7179073002198592,
      "grad_norm": 0.032072387635707855,
      "learning_rate": 3.8534883130104934e-05,
      "loss": 0.027,
      "step": 4000
    },
    {
      "epoch": 0.7997487324449231,
      "eval_cer": 0.7718523934618883,
      "eval_loss": 0.01841309294104576,
      "eval_runtime": 436.0956,
      "eval_samples_per_second": 406.037,
      "eval_steps_per_second": 1.059,
      "step": 4456
    },
    {
      "epoch": 0.8076457127473415,
      "grad_norm": 0.0274814460426569,
      "learning_rate": 3.2152071176988856e-05,
      "loss": 0.0248,
      "step": 4500
    },
    {
      "epoch": 0.8973841252748239,
      "grad_norm": 0.025461740791797638,
      "learning_rate": 2.520441835335607e-05,
      "loss": 0.0232,
      "step": 5000
    },
    {
      "epoch": 0.9871225378023063,
      "grad_norm": 0.026277432218194008,
      "learning_rate": 1.8240621417367316e-05,
      "loss": 0.0222,
      "step": 5500
    },
    {
      "epoch": 0.9996859155561538,
      "eval_cer": 0.7717538371980056,
      "eval_loss": 0.016575267538428307,
      "eval_runtime": 437.5161,
      "eval_samples_per_second": 404.719,
      "eval_steps_per_second": 1.056,
      "step": 5570
    }
  ],
  "logging_steps": 500,
  "max_steps": 5571,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 1114,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 6.522138159281603e+17,
  "train_batch_size": 384,
  "trial_name": null,
  "trial_params": null
}