trainer_state.json

{
  "best_metric": 0.4165333333333333,
  "best_model_checkpoint": "./vit_finetuned_models_dataset/CIFAR100/50_from_100/facebook_dino-vitb16/model_idx_0503/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": 26.358625411987305,
      "learning_rate": 0.00028549422332891283,
      "loss": 3.2742,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.19573333333333334,
      "eval_loss": 2.9345805644989014,
      "eval_runtime": 10.3609,
      "eval_samples_per_second": 361.936,
      "eval_steps_per_second": 5.694,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": 43.879234313964844,
      "learning_rate": 0.00024415433369325305,
      "loss": 2.7211,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.2536,
      "eval_loss": 2.6903817653656006,
      "eval_runtime": 10.1196,
      "eval_samples_per_second": 370.568,
      "eval_steps_per_second": 5.83,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 31.555166244506836,
      "learning_rate": 0.0001841660235425207,
      "loss": 2.4701,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.32666666666666666,
      "eval_loss": 2.4302573204040527,
      "eval_runtime": 10.1,
      "eval_samples_per_second": 371.286,
      "eval_steps_per_second": 5.842,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": 23.157106399536133,
      "learning_rate": 0.00011741071341041153,
      "loss": 2.2143,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.37146666666666667,
      "eval_loss": 2.236201524734497,
      "eval_runtime": 10.4615,
      "eval_samples_per_second": 358.459,
      "eval_steps_per_second": 5.64,
      "step": 1332
    },
    {
      "epoch": 5.0,
      "grad_norm": 39.43628692626953,
      "learning_rate": 5.711011116874128e-05,
      "loss": 1.9196,
      "step": 1665
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.4024,
      "eval_loss": 2.148569345474243,
      "eval_runtime": 10.1338,
      "eval_samples_per_second": 370.049,
      "eval_steps_per_second": 5.822,
      "step": 1665
    },
    {
      "epoch": 6.0,
      "grad_norm": 62.00448226928711,
      "learning_rate": 1.5207490629826985e-05,
      "loss": 1.5588,
      "step": 1998
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.4149333333333333,
      "eval_loss": 2.131460666656494,
      "eval_runtime": 10.1517,
      "eval_samples_per_second": 369.398,
      "eval_steps_per_second": 5.812,
      "step": 1998
    },
    {
      "epoch": 7.0,
      "grad_norm": Infinity,
      "learning_rate": 3.4057661676356597e-09,
      "loss": 1.2312,
      "step": 2331
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.4165333333333333,
      "eval_loss": 2.272395610809326,
      "eval_runtime": 10.1959,
      "eval_samples_per_second": 367.796,
      "eval_steps_per_second": 5.787,
      "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
}