trainer_state.json

{
  "best_metric": 0.6426666666666667,
  "best_model_checkpoint": "./vit_finetuned_models_dataset/CIFAR100/50_from_100/facebook_vit-mae-base/model_idx_0690/checkpoints/checkpoint-1998",
  "epoch": 6.0,
  "eval_steps": 500,
  "global_step": 1998,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": Infinity,
      "learning_rate": 0.0005,
      "loss": 2.9504,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.3416,
      "eval_loss": 2.3998796939849854,
      "eval_runtime": 6.1812,
      "eval_samples_per_second": 606.676,
      "eval_steps_per_second": 9.545,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": Infinity,
      "learning_rate": 0.0005,
      "loss": 2.023,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.4928,
      "eval_loss": 1.7860721349716187,
      "eval_runtime": 6.1143,
      "eval_samples_per_second": 613.316,
      "eval_steps_per_second": 9.649,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 6.22600793838501,
      "learning_rate": 0.0005,
      "loss": 1.4413,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.5637333333333333,
      "eval_loss": 1.4792795181274414,
      "eval_runtime": 6.136,
      "eval_samples_per_second": 611.148,
      "eval_steps_per_second": 9.615,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": 9.439937591552734,
      "learning_rate": 0.0005,
      "loss": 1.0592,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.5901333333333333,
      "eval_loss": 1.471636176109314,
      "eval_runtime": 6.4508,
      "eval_samples_per_second": 581.327,
      "eval_steps_per_second": 9.146,
      "step": 1332
    },
    {
      "epoch": 5.0,
      "grad_norm": 18.574417114257812,
      "learning_rate": 0.0005,
      "loss": 0.7655,
      "step": 1665
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.6312,
      "eval_loss": 1.2981716394424438,
      "eval_runtime": 6.6957,
      "eval_samples_per_second": 560.06,
      "eval_steps_per_second": 8.812,
      "step": 1665
    },
    {
      "epoch": 6.0,
      "grad_norm": 5.229493618011475,
      "learning_rate": 0.0005,
      "loss": 0.5449,
      "step": 1998
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.6426666666666667,
      "eval_loss": 1.3248109817504883,
      "eval_runtime": 6.1184,
      "eval_samples_per_second": 612.901,
      "eval_steps_per_second": 9.643,
      "step": 1998
    }
  ],
  "logging_steps": 500,
  "max_steps": 2664,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 8,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 9.88447923735552e+18,
  "train_batch_size": 64,
  "trial_name": null,
  "trial_params": null
}