trainer_state.json

{
  "best_metric": 0.2626006603240967,
  "best_model_checkpoint": "../models/b_bl_asc1k/checkpoint-378",
  "epoch": 6.0,
  "eval_steps": 126,
  "global_step": 378,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.015873015873015872,
      "grad_norm": 8.237451553344727,
      "learning_rate": 1.5873015873015872e-05,
      "loss": 4.6114,
      "step": 1
    },
    {
      "epoch": 1.0,
      "grad_norm": 1.6461960077285767,
      "learning_rate": 0.001,
      "loss": 2.1089,
      "step": 63
    },
    {
      "epoch": 2.0,
      "grad_norm": 2.783024787902832,
      "learning_rate": 0.0008888888888888888,
      "loss": 0.5071,
      "step": 126
    },
    {
      "epoch": 2.0,
      "eval_loss": 0.4069588780403137,
      "eval_runtime": 125.8708,
      "eval_samples_per_second": 1.446,
      "eval_steps_per_second": 0.183,
      "step": 126
    },
    {
      "epoch": 3.0,
      "grad_norm": 1.3634392023086548,
      "learning_rate": 0.0007777777777777778,
      "loss": 0.3171,
      "step": 189
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.9046511054039001,
      "learning_rate": 0.0006666666666666666,
      "loss": 0.2359,
      "step": 252
    },
    {
      "epoch": 4.0,
      "eval_loss": 0.29633015394210815,
      "eval_runtime": 64.9079,
      "eval_samples_per_second": 2.804,
      "eval_steps_per_second": 0.354,
      "step": 252
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.7224295735359192,
      "learning_rate": 0.0005555555555555556,
      "loss": 0.1861,
      "step": 315
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.9205335378646851,
      "learning_rate": 0.0004444444444444444,
      "loss": 0.149,
      "step": 378
    },
    {
      "epoch": 6.0,
      "eval_loss": 0.2626006603240967,
      "eval_runtime": 64.5046,
      "eval_samples_per_second": 2.822,
      "eval_steps_per_second": 0.357,
      "step": 378
    }
  ],
  "logging_steps": 63,
  "max_steps": 630,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 10,
  "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": 3.9595474944e+17,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}