trainer_state.json

{
  "best_metric": 0.49306666666666665,
  "best_model_checkpoint": "./vit_finetuned_models_dataset/CIFAR100/50_from_100/facebook_dino-vitb16/model_idx_0335/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": 23.721229553222656,
      "learning_rate": 0.0002910611396268257,
      "loss": 3.2957,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.21866666666666668,
      "eval_loss": 2.9166531562805176,
      "eval_runtime": 4.2978,
      "eval_samples_per_second": 872.548,
      "eval_steps_per_second": 13.728,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": 19.95889663696289,
      "learning_rate": 0.0002651085536196122,
      "loss": 2.737,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.2618666666666667,
      "eval_loss": 2.774722099304199,
      "eval_runtime": 4.4467,
      "eval_samples_per_second": 843.325,
      "eval_steps_per_second": 13.268,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 33.750244140625,
      "learning_rate": 0.00022527217722455154,
      "loss": 2.4704,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.33813333333333334,
      "eval_loss": 2.4365029335021973,
      "eval_runtime": 6.2058,
      "eval_samples_per_second": 604.273,
      "eval_steps_per_second": 9.507,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": 26.89750099182129,
      "learning_rate": 0.00017635686535718819,
      "loss": 2.2189,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.3794666666666667,
      "eval_loss": 2.25345516204834,
      "eval_runtime": 4.2668,
      "eval_samples_per_second": 878.872,
      "eval_steps_per_second": 13.828,
      "step": 1332
    },
    {
      "epoch": 5.0,
      "grad_norm": 27.6105899810791,
      "learning_rate": 0.00012426252654184345,
      "loss": 1.9743,
      "step": 1665
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.4232,
      "eval_loss": 2.069655656814575,
      "eval_runtime": 6.1325,
      "eval_samples_per_second": 611.491,
      "eval_steps_per_second": 9.621,
      "step": 1665
    },
    {
      "epoch": 6.0,
      "grad_norm": 34.59181594848633,
      "learning_rate": 7.527250687020608e-05,
      "loss": 1.7332,
      "step": 1998
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.46426666666666666,
      "eval_loss": 1.952174186706543,
      "eval_runtime": 4.5929,
      "eval_samples_per_second": 816.476,
      "eval_steps_per_second": 12.846,
      "step": 1998
    },
    {
      "epoch": 7.0,
      "grad_norm": Infinity,
      "learning_rate": 3.5397110943979596e-05,
      "loss": 1.4633,
      "step": 2331
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.4786666666666667,
      "eval_loss": 1.9194862842559814,
      "eval_runtime": 4.3083,
      "eval_samples_per_second": 870.411,
      "eval_steps_per_second": 13.694,
      "step": 2331
    },
    {
      "epoch": 8.0,
      "grad_norm": 30.46764373779297,
      "learning_rate": 9.20813811843163e-06,
      "loss": 1.2094,
      "step": 2664
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.4842666666666667,
      "eval_loss": 1.9729022979736328,
      "eval_runtime": 4.2601,
      "eval_samples_per_second": 880.26,
      "eval_steps_per_second": 13.849,
      "step": 2664
    },
    {
      "epoch": 9.0,
      "grad_norm": 23.135236740112305,
      "learning_rate": 7.417023831113311e-10,
      "loss": 1.0419,
      "step": 2997
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.49306666666666665,
      "eval_loss": 2.0152900218963623,
      "eval_runtime": 4.4313,
      "eval_samples_per_second": 846.259,
      "eval_steps_per_second": 13.314,
      "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
}