trainer_state.json

{
  "best_metric": 0.9042666666666667,
  "best_model_checkpoint": "./vit_finetuned_models_dataset/CIFAR100/50_from_100/facebook_dino-vitb16/model_idx_0146/checkpoints/checkpoint-2331",
  "epoch": 7.0,
  "eval_steps": 500,
  "global_step": 2331,
  "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": 8.580008580008581e-05,
      "loss": 0.9151,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8338666666666666,
      "eval_loss": 0.5448431372642517,
      "eval_runtime": 12.3612,
      "eval_samples_per_second": 303.368,
      "eval_steps_per_second": 4.773,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.2614489793777466,
      "learning_rate": 7.151437151437152e-05,
      "loss": 0.3368,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8322666666666667,
      "eval_loss": 0.5604428648948669,
      "eval_runtime": 11.8774,
      "eval_samples_per_second": 315.725,
      "eval_steps_per_second": 4.967,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.19073085486888885,
      "learning_rate": 5.7228657228657234e-05,
      "loss": 0.1689,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8621333333333333,
      "eval_loss": 0.5110751986503601,
      "eval_runtime": 11.7951,
      "eval_samples_per_second": 317.929,
      "eval_steps_per_second": 5.002,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.9606668949127197,
      "learning_rate": 4.294294294294294e-05,
      "loss": 0.0962,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8722666666666666,
      "eval_loss": 0.4860284626483917,
      "eval_runtime": 11.8503,
      "eval_samples_per_second": 316.448,
      "eval_steps_per_second": 4.979,
      "step": 1332
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.001516719232313335,
      "learning_rate": 2.8657228657228656e-05,
      "loss": 0.043,
      "step": 1665
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.8757333333333334,
      "eval_loss": 0.47397297620773315,
      "eval_runtime": 11.9493,
      "eval_samples_per_second": 313.827,
      "eval_steps_per_second": 4.938,
      "step": 1665
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.0349971167743206,
      "learning_rate": 1.4371514371514371e-05,
      "loss": 0.0152,
      "step": 1998
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.9008,
      "eval_loss": 0.4094306230545044,
      "eval_runtime": 11.8394,
      "eval_samples_per_second": 316.738,
      "eval_steps_per_second": 4.983,
      "step": 1998
    },
    {
      "epoch": 7.0,
      "grad_norm": 0.003355839056894183,
      "learning_rate": 8.580008580008581e-08,
      "loss": 0.0034,
      "step": 2331
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.9042666666666667,
      "eval_loss": 0.39918360114097595,
      "eval_runtime": 11.9635,
      "eval_samples_per_second": 313.454,
      "eval_steps_per_second": 4.932,
      "step": 2331
    }
  ],
  "logging_steps": 500,
  "max_steps": 2331,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 7,
  "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": 1.153189244358144e+19,
  "train_batch_size": 64,
  "trial_name": null,
  "trial_params": null
}