trainer_state.json

{
  "best_metric": 0.9090666666666667,
  "best_model_checkpoint": "./vit_finetuned_models_dataset/CIFAR100/50_from_100/facebook_dino-vitb16/model_idx_0863/checkpoints/checkpoint-2664",
  "epoch": 8.0,
  "eval_steps": 500,
  "global_step": 2664,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 45.02859115600586,
      "learning_rate": 9.623897705919314e-05,
      "loss": 0.9935,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.7997333333333333,
      "eval_loss": 0.6757562756538391,
      "eval_runtime": 11.9455,
      "eval_samples_per_second": 313.927,
      "eval_steps_per_second": 4.939,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": 59.14358901977539,
      "learning_rate": 8.543862808774192e-05,
      "loss": 0.4062,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8130666666666667,
      "eval_loss": 0.6657377481460571,
      "eval_runtime": 11.3166,
      "eval_samples_per_second": 331.373,
      "eval_steps_per_second": 5.214,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.059999801218509674,
      "learning_rate": 6.924306924189557e-05,
      "loss": 0.2337,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8608,
      "eval_loss": 0.4907033443450928,
      "eval_runtime": 10.9685,
      "eval_samples_per_second": 341.888,
      "eval_steps_per_second": 5.379,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.01627151295542717,
      "learning_rate": 5.011792754282769e-05,
      "loss": 0.1258,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8682666666666666,
      "eval_loss": 0.5155117511749268,
      "eval_runtime": 10.5487,
      "eval_samples_per_second": 355.493,
      "eval_steps_per_second": 5.593,
      "step": 1332
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.03148584067821503,
      "learning_rate": 3.0974832444380136e-05,
      "loss": 0.0684,
      "step": 1665
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.8845333333333333,
      "eval_loss": 0.4530421793460846,
      "eval_runtime": 10.5848,
      "eval_samples_per_second": 354.282,
      "eval_steps_per_second": 5.574,
      "step": 1665
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.003269389970228076,
      "learning_rate": 1.472814664270234e-05,
      "loss": 0.0249,
      "step": 1998
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.9026666666666666,
      "eval_loss": 0.40313228964805603,
      "eval_runtime": 10.8551,
      "eval_samples_per_second": 345.46,
      "eval_steps_per_second": 5.435,
      "step": 1998
    },
    {
      "epoch": 7.0,
      "grad_norm": 0.006605663802474737,
      "learning_rate": 3.851280774526222e-06,
      "loss": 0.007,
      "step": 2331
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.9069333333333334,
      "eval_loss": 0.39270541071891785,
      "eval_runtime": 10.5586,
      "eval_samples_per_second": 355.162,
      "eval_steps_per_second": 5.588,
      "step": 2331
    },
    {
      "epoch": 8.0,
      "grad_norm": 0.013314406387507915,
      "learning_rate": 1.3906924697382196e-10,
      "loss": 0.002,
      "step": 2664
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.9090666666666667,
      "eval_loss": 0.38259774446487427,
      "eval_runtime": 12.9891,
      "eval_samples_per_second": 288.705,
      "eval_steps_per_second": 4.542,
      "step": 2664
    }
  ],
  "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": true
      },
      "attributes": {}
    }
  },
  "total_flos": 1.317930564980736e+19,
  "train_batch_size": 64,
  "trial_name": null,
  "trial_params": null
}