trainer_state.json

{
  "best_metric": 0.8928,
  "best_model_checkpoint": "./vit_finetuned_models_dataset/CIFAR100/50_from_100/facebook_vit-mae-base/model_idx_0628/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": 2.0312647819519043,
      "learning_rate": 9.698463103929542e-05,
      "loss": 1.5144,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.7736,
      "eval_loss": 0.7882294058799744,
      "eval_runtime": 6.8092,
      "eval_samples_per_second": 550.727,
      "eval_steps_per_second": 8.665,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": Infinity,
      "learning_rate": 8.833589107668211e-05,
      "loss": 0.5244,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8266666666666667,
      "eval_loss": 0.5681686401367188,
      "eval_runtime": 7.351,
      "eval_samples_per_second": 510.135,
      "eval_steps_per_second": 8.026,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.33514222502708435,
      "learning_rate": 7.504537663092833e-05,
      "loss": 0.3018,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8541333333333333,
      "eval_loss": 0.5153558254241943,
      "eval_runtime": 6.5912,
      "eval_samples_per_second": 568.936,
      "eval_steps_per_second": 8.951,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.9088128805160522,
      "learning_rate": 5.873402013309226e-05,
      "loss": 0.1761,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8581333333333333,
      "eval_loss": 0.49346211552619934,
      "eval_runtime": 7.1618,
      "eval_samples_per_second": 523.609,
      "eval_steps_per_second": 8.238,
      "step": 1332
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.7470325231552124,
      "learning_rate": 4.136921190679637e-05,
      "loss": 0.1044,
      "step": 1665
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.8693333333333333,
      "eval_loss": 0.45761653780937195,
      "eval_runtime": 6.6512,
      "eval_samples_per_second": 563.81,
      "eval_steps_per_second": 8.871,
      "step": 1665
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.030977608636021614,
      "learning_rate": 2.5045404101407076e-05,
      "loss": 0.0508,
      "step": 1998
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.8864,
      "eval_loss": 0.44482100009918213,
      "eval_runtime": 6.6971,
      "eval_samples_per_second": 559.943,
      "eval_steps_per_second": 8.81,
      "step": 1998
    },
    {
      "epoch": 7.0,
      "grad_norm": 0.09944266825914383,
      "learning_rate": 1.1731488851999511e-05,
      "loss": 0.0226,
      "step": 2331
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.8912,
      "eval_loss": 0.433264821767807,
      "eval_runtime": 6.7872,
      "eval_samples_per_second": 552.512,
      "eval_steps_per_second": 8.693,
      "step": 2331
    },
    {
      "epoch": 8.0,
      "grad_norm": 0.2899631857872009,
      "learning_rate": 3.033320830114261e-06,
      "loss": 0.0114,
      "step": 2664
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.8917333333333334,
      "eval_loss": 0.43743112683296204,
      "eval_runtime": 6.681,
      "eval_samples_per_second": 561.292,
      "eval_steps_per_second": 8.831,
      "step": 2664
    },
    {
      "epoch": 9.0,
      "grad_norm": 0.03425300121307373,
      "learning_rate": 2.7470478758395702e-11,
      "loss": 0.0072,
      "step": 2997
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.8928,
      "eval_loss": 0.4280654489994049,
      "eval_runtime": 6.6252,
      "eval_samples_per_second": 566.024,
      "eval_steps_per_second": 8.905,
      "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
}