trainer_state.json

{
  "best_metric": 0.15153372287750244,
  "best_model_checkpoint": "../models/b_ll_asc1k-2/checkpoint-378",
  "epoch": 5.1,
  "eval_steps": 126,
  "global_step": 378,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0015873015873015873,
      "grad_norm": 1.7664669752120972,
      "learning_rate": 1e-05,
      "loss": 2.9008,
      "step": 1
    },
    {
      "epoch": 0.1,
      "grad_norm": 1.1034066677093506,
      "learning_rate": 0.00063,
      "loss": 1.2295,
      "step": 63
    },
    {
      "epoch": 1.1,
      "grad_norm": 0.6373379230499268,
      "learning_rate": 0.0009509433962264152,
      "loss": 0.2402,
      "step": 126
    },
    {
      "epoch": 1.1,
      "eval_loss": 0.20612144470214844,
      "eval_runtime": 25.198,
      "eval_samples_per_second": 7.223,
      "eval_steps_per_second": 0.913,
      "step": 126
    },
    {
      "epoch": 2.1,
      "grad_norm": 0.5278385281562805,
      "learning_rate": 0.0008320754716981132,
      "loss": 0.1486,
      "step": 189
    },
    {
      "epoch": 3.1,
      "grad_norm": 0.562412440776825,
      "learning_rate": 0.0007132075471698113,
      "loss": 0.1,
      "step": 252
    },
    {
      "epoch": 3.1,
      "eval_loss": 0.1705094277858734,
      "eval_runtime": 24.6952,
      "eval_samples_per_second": 7.37,
      "eval_steps_per_second": 0.931,
      "step": 252
    },
    {
      "epoch": 4.1,
      "grad_norm": 0.3925608992576599,
      "learning_rate": 0.0005943396226415095,
      "loss": 0.0655,
      "step": 315
    },
    {
      "epoch": 5.1,
      "grad_norm": 0.39186909794807434,
      "learning_rate": 0.0004754716981132076,
      "loss": 0.0411,
      "step": 378
    },
    {
      "epoch": 5.1,
      "eval_loss": 0.15153372287750244,
      "eval_runtime": 24.8181,
      "eval_samples_per_second": 7.333,
      "eval_steps_per_second": 0.927,
      "step": 378
    }
  ],
  "logging_steps": 63,
  "max_steps": 630,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 9223372036854775807,
  "save_steps": 126,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 1.28296230912e+19,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}