trainer_state.json

{
  "best_metric": 0.8706666666666667,
  "best_model_checkpoint": "./vit_finetuned_models_dataset/CIFAR100/50_from_100/facebook_vit-mae-base/model_idx_0797/checkpoints/checkpoint-2997",
  "epoch": 9.0,
  "eval_steps": 500,
  "global_step": 2997,
  "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": 2.9100759384327822e-05,
      "loss": 2.3644,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.7189333333333333,
      "eval_loss": 1.1851471662521362,
      "eval_runtime": 6.1717,
      "eval_samples_per_second": 607.609,
      "eval_steps_per_second": 9.56,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": Infinity,
      "learning_rate": 2.6510855361961218e-05,
      "loss": 0.8085,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8109333333333333,
      "eval_loss": 0.6939958333969116,
      "eval_runtime": 6.0706,
      "eval_samples_per_second": 617.731,
      "eval_steps_per_second": 9.719,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": Infinity,
      "learning_rate": 2.2540814184580827e-05,
      "loss": 0.4112,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8536,
      "eval_loss": 0.5211882591247559,
      "eval_runtime": 6.5221,
      "eval_samples_per_second": 574.965,
      "eval_steps_per_second": 9.046,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": 98.79678344726562,
      "learning_rate": 1.7651164135367127e-05,
      "loss": 0.2253,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8586666666666667,
      "eval_loss": 0.48992305994033813,
      "eval_runtime": 6.1122,
      "eval_samples_per_second": 613.525,
      "eval_steps_per_second": 9.653,
      "step": 1332
    },
    {
      "epoch": 5.0,
      "grad_norm": 129.96434020996094,
      "learning_rate": 1.2441744564412656e-05,
      "loss": 0.1154,
      "step": 1665
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.8629333333333333,
      "eval_loss": 0.46051424741744995,
      "eval_runtime": 6.1078,
      "eval_samples_per_second": 613.969,
      "eval_steps_per_second": 9.66,
      "step": 1665
    },
    {
      "epoch": 6.0,
      "grad_norm": 7.945436954498291,
      "learning_rate": 7.5408883548191384e-06,
      "loss": 0.0538,
      "step": 1998
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.8674666666666667,
      "eval_loss": 0.4722992777824402,
      "eval_runtime": 6.1367,
      "eval_samples_per_second": 611.075,
      "eval_steps_per_second": 9.614,
      "step": 1998
    },
    {
      "epoch": 7.0,
      "grad_norm": 2.178852081298828,
      "learning_rate": 3.5397110943979598e-06,
      "loss": 0.0265,
      "step": 2331
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.868,
      "eval_loss": 0.47248902916908264,
      "eval_runtime": 6.0978,
      "eval_samples_per_second": 614.978,
      "eval_steps_per_second": 9.676,
      "step": 2331
    },
    {
      "epoch": 8.0,
      "grad_norm": 2.1545047760009766,
      "learning_rate": 9.20813811843163e-07,
      "loss": 0.0169,
      "step": 2664
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.8704,
      "eval_loss": 0.47840356826782227,
      "eval_runtime": 6.1271,
      "eval_samples_per_second": 612.035,
      "eval_steps_per_second": 9.629,
      "step": 2664
    },
    {
      "epoch": 9.0,
      "grad_norm": 0.1360931098461151,
      "learning_rate": 7.417023831113313e-11,
      "loss": 0.0139,
      "step": 2997
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.8706666666666667,
      "eval_loss": 0.4790334105491638,
      "eval_runtime": 6.6183,
      "eval_samples_per_second": 566.611,
      "eval_steps_per_second": 8.915,
      "step": 2997
    }
  ],
  "logging_steps": 500,
  "max_steps": 2997,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 9,
  "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": 1.482671885603328e+19,
  "train_batch_size": 64,
  "trial_name": null,
  "trial_params": null
}