trainer_state.json

{
  "best_metric": 0.8610666666666666,
  "best_model_checkpoint": "./vit_finetuned_models_dataset/CIFAR100/50_from_100/facebook_vit-mae-base/model_idx_0264/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": 0.0002910075938432782,
      "loss": 2.431,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.6496,
      "eval_loss": 1.2076555490493774,
      "eval_runtime": 5.9791,
      "eval_samples_per_second": 627.189,
      "eval_steps_per_second": 9.868,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": Infinity,
      "learning_rate": 0.0002651085536196122,
      "loss": 0.8945,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.7656,
      "eval_loss": 0.7904889583587646,
      "eval_runtime": 5.856,
      "eval_samples_per_second": 640.368,
      "eval_steps_per_second": 10.075,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": Infinity,
      "learning_rate": 0.00022540814184580826,
      "loss": 0.5114,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8093333333333333,
      "eval_loss": 0.650262176990509,
      "eval_runtime": 6.2992,
      "eval_samples_per_second": 595.316,
      "eval_steps_per_second": 9.366,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": 3.902454137802124,
      "learning_rate": 0.00017651164135367125,
      "loss": 0.2882,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8181333333333334,
      "eval_loss": 0.5918785333633423,
      "eval_runtime": 6.3739,
      "eval_samples_per_second": 588.333,
      "eval_steps_per_second": 9.256,
      "step": 1332
    },
    {
      "epoch": 5.0,
      "grad_norm": 1.3055394887924194,
      "learning_rate": 0.00012441744564412654,
      "loss": 0.1517,
      "step": 1665
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.8378666666666666,
      "eval_loss": 0.5970044136047363,
      "eval_runtime": 6.0051,
      "eval_samples_per_second": 624.467,
      "eval_steps_per_second": 9.825,
      "step": 1665
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.08424951136112213,
      "learning_rate": 7.540888354819138e-05,
      "loss": 0.0656,
      "step": 1998
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.8453333333333334,
      "eval_loss": 0.5983319282531738,
      "eval_runtime": 6.1291,
      "eval_samples_per_second": 611.832,
      "eval_steps_per_second": 9.626,
      "step": 1998
    },
    {
      "epoch": 7.0,
      "grad_norm": 0.0152902752161026,
      "learning_rate": 3.5397110943979596e-05,
      "loss": 0.0198,
      "step": 2331
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.8528,
      "eval_loss": 0.5852423310279846,
      "eval_runtime": 5.8689,
      "eval_samples_per_second": 638.959,
      "eval_steps_per_second": 10.053,
      "step": 2331
    },
    {
      "epoch": 8.0,
      "grad_norm": 0.135624498128891,
      "learning_rate": 9.20813811843163e-06,
      "loss": 0.0082,
      "step": 2664
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.8594666666666667,
      "eval_loss": 0.5606937408447266,
      "eval_runtime": 5.8107,
      "eval_samples_per_second": 645.362,
      "eval_steps_per_second": 10.154,
      "step": 2664
    },
    {
      "epoch": 9.0,
      "grad_norm": 0.2623247802257538,
      "learning_rate": 7.417023831113311e-10,
      "loss": 0.0036,
      "step": 2997
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.8610666666666666,
      "eval_loss": 0.5646731853485107,
      "eval_runtime": 5.8506,
      "eval_samples_per_second": 640.956,
      "eval_steps_per_second": 10.084,
      "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
}