trainer_state.json

{
  "best_metric": 0.4448,
  "best_model_checkpoint": "./vit_finetuned_models_dataset/CIFAR100/50_from_100/facebook_dino-vitb16/model_idx_0396/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": 0.0002887841425059038,
      "loss": 3.3818,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.21626666666666666,
      "eval_loss": 3.012765407562256,
      "eval_runtime": 5.0996,
      "eval_samples_per_second": 735.354,
      "eval_steps_per_second": 11.57,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": 18.01856231689453,
      "learning_rate": 0.0002564405961135417,
      "loss": 2.8554,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.2541333333333333,
      "eval_loss": 2.7979013919830322,
      "eval_runtime": 4.3148,
      "eval_samples_per_second": 869.1,
      "eval_steps_per_second": 13.674,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 35.2623291015625,
      "learning_rate": 0.0002078924338492993,
      "loss": 2.5919,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.2965333333333333,
      "eval_loss": 2.6181223392486572,
      "eval_runtime": 4.2986,
      "eval_samples_per_second": 872.369,
      "eval_steps_per_second": 13.725,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": 22.049970626831055,
      "learning_rate": 0.00015053067332772077,
      "loss": 2.3414,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.3658666666666667,
      "eval_loss": 2.3733301162719727,
      "eval_runtime": 6.1597,
      "eval_samples_per_second": 608.796,
      "eval_steps_per_second": 9.578,
      "step": 1332
    },
    {
      "epoch": 5.0,
      "grad_norm": 21.74603843688965,
      "learning_rate": 9.308812260256249e-05,
      "loss": 2.0824,
      "step": 1665
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.4013333333333333,
      "eval_loss": 2.2389073371887207,
      "eval_runtime": 4.973,
      "eval_samples_per_second": 754.065,
      "eval_steps_per_second": 11.864,
      "step": 1665
    },
    {
      "epoch": 6.0,
      "grad_norm": 33.30829620361328,
      "learning_rate": 4.4309889303710674e-05,
      "loss": 1.8223,
      "step": 1998
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.4202666666666667,
      "eval_loss": 2.182614326477051,
      "eval_runtime": 4.4063,
      "eval_samples_per_second": 851.049,
      "eval_steps_per_second": 13.39,
      "step": 1998
    },
    {
      "epoch": 7.0,
      "grad_norm": 26.433942794799805,
      "learning_rate": 1.1622017275129708e-05,
      "loss": 1.567,
      "step": 2331
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.4448,
      "eval_loss": 2.1260054111480713,
      "eval_runtime": 4.2921,
      "eval_samples_per_second": 873.704,
      "eval_steps_per_second": 13.746,
      "step": 2331
    }
  ],
  "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": 1.153189244358144e+19,
  "train_batch_size": 64,
  "trial_name": null,
  "trial_params": null
}