trainer_state.json

{
  "best_metric": 1.1339149475097656,
  "best_model_checkpoint": "../models/a_bl1kd10-re/checkpoint-440",
  "epoch": 3.1,
  "eval_steps": 220,
  "global_step": 440,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.1,
      "grad_norm": 1.454114317893982,
      "learning_rate": 0.000990909090909091,
      "loss": 2.5139,
      "step": 110
    },
    {
      "epoch": 1.1,
      "grad_norm": 1.0413837432861328,
      "learning_rate": 0.0008898989898989899,
      "loss": 1.1605,
      "step": 220
    },
    {
      "epoch": 1.1,
      "eval_loss": 1.3539080619812012,
      "eval_runtime": 118.9056,
      "eval_samples_per_second": 2.944,
      "eval_steps_per_second": 0.185,
      "step": 220
    },
    {
      "epoch": 2.1,
      "grad_norm": 1.1177339553833008,
      "learning_rate": 0.0007787878787878788,
      "loss": 0.8597,
      "step": 330
    },
    {
      "epoch": 3.1,
      "grad_norm": 1.0981805324554443,
      "learning_rate": 0.0006676767676767676,
      "loss": 0.7193,
      "step": 440
    },
    {
      "epoch": 3.1,
      "eval_loss": 1.1339149475097656,
      "eval_runtime": 118.0745,
      "eval_samples_per_second": 2.964,
      "eval_steps_per_second": 0.186,
      "step": 440
    }
  ],
  "logging_steps": 110,
  "max_steps": 1100,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 9223372036854775807,
  "save_steps": 220,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 1.84778883072e+18,
  "train_batch_size": 64,
  "trial_name": null,
  "trial_params": null
}