trainer_state.json

{
  "best_metric": 0.9921259842519685,
  "best_model_checkpoint": "convnext-xlarge-224-22k-1k-v14/checkpoint-18",
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 27,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "eval_accuracy": 0.9606299212598425,
      "eval_loss": 0.1334373652935028,
      "eval_runtime": 2.5798,
      "eval_samples_per_second": 49.229,
      "eval_steps_per_second": 1.551,
      "step": 9
    },
    {
      "epoch": 1.1111111111111112,
      "grad_norm": 135.50009155273438,
      "learning_rate": 3.541666666666667e-05,
      "loss": 0.2411,
      "step": 10
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9921259842519685,
      "eval_loss": 0.03504885360598564,
      "eval_runtime": 1.5915,
      "eval_samples_per_second": 79.8,
      "eval_steps_per_second": 2.513,
      "step": 18
    },
    {
      "epoch": 2.2222222222222223,
      "grad_norm": 9.684277534484863,
      "learning_rate": 1.4583333333333335e-05,
      "loss": 0.0722,
      "step": 20
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9921259842519685,
      "eval_loss": 0.01645192690193653,
      "eval_runtime": 1.6699,
      "eval_samples_per_second": 76.05,
      "eval_steps_per_second": 2.395,
      "step": 27
    },
    {
      "epoch": 3.0,
      "step": 27,
      "total_flos": 1.0772707724270715e+18,
      "train_loss": 0.12385351238427339,
      "train_runtime": 162.1968,
      "train_samples_per_second": 21.122,
      "train_steps_per_second": 0.166
    }
  ],
  "logging_steps": 10,
  "max_steps": 27,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 1.0772707724270715e+18,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}