trainer_state.json

{
  "best_metric": 0.9982142857142857,
  "best_model_checkpoint": "vit-large-patch16-224-dungeon-geo-morphs-0-4-30Nov24-006/checkpoint-60",
  "epoch": 35.45454545454545,
  "eval_steps": 10,
  "global_step": 80,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 4.363636363636363,
      "grad_norm": 24.64277458190918,
      "learning_rate": 9.861111111111112e-06,
      "loss": 1.5574,
      "step": 10
    },
    {
      "epoch": 4.363636363636363,
      "eval_accuracy": 0.6017857142857143,
      "eval_loss": 1.2378439903259277,
      "eval_runtime": 7.5027,
      "eval_samples_per_second": 74.64,
      "eval_steps_per_second": 9.33,
      "step": 10
    },
    {
      "epoch": 8.727272727272727,
      "grad_norm": 23.63115119934082,
      "learning_rate": 8.472222222222223e-06,
      "loss": 0.9431,
      "step": 20
    },
    {
      "epoch": 8.727272727272727,
      "eval_accuracy": 0.8892857142857142,
      "eval_loss": 0.6701930165290833,
      "eval_runtime": 7.7778,
      "eval_samples_per_second": 72.0,
      "eval_steps_per_second": 9.0,
      "step": 20
    },
    {
      "epoch": 13.272727272727273,
      "grad_norm": 14.12543773651123,
      "learning_rate": 7.083333333333335e-06,
      "loss": 0.4807,
      "step": 30
    },
    {
      "epoch": 13.272727272727273,
      "eval_accuracy": 0.95,
      "eval_loss": 0.38032934069633484,
      "eval_runtime": 7.2067,
      "eval_samples_per_second": 77.706,
      "eval_steps_per_second": 9.713,
      "step": 30
    },
    {
      "epoch": 17.636363636363637,
      "grad_norm": 13.682252883911133,
      "learning_rate": 5.694444444444445e-06,
      "loss": 0.2673,
      "step": 40
    },
    {
      "epoch": 17.636363636363637,
      "eval_accuracy": 0.9732142857142857,
      "eval_loss": 0.22460468113422394,
      "eval_runtime": 8.2781,
      "eval_samples_per_second": 67.648,
      "eval_steps_per_second": 8.456,
      "step": 40
    },
    {
      "epoch": 22.181818181818183,
      "grad_norm": 21.752925872802734,
      "learning_rate": 4.305555555555556e-06,
      "loss": 0.1501,
      "step": 50
    },
    {
      "epoch": 22.181818181818183,
      "eval_accuracy": 0.975,
      "eval_loss": 0.15731294453144073,
      "eval_runtime": 7.5746,
      "eval_samples_per_second": 73.931,
      "eval_steps_per_second": 9.241,
      "step": 50
    },
    {
      "epoch": 26.545454545454547,
      "grad_norm": 8.0554838180542,
      "learning_rate": 2.916666666666667e-06,
      "loss": 0.0962,
      "step": 60
    },
    {
      "epoch": 26.545454545454547,
      "eval_accuracy": 0.9982142857142857,
      "eval_loss": 0.10373682528734207,
      "eval_runtime": 8.3878,
      "eval_samples_per_second": 66.764,
      "eval_steps_per_second": 8.346,
      "step": 60
    },
    {
      "epoch": 31.09090909090909,
      "grad_norm": 6.769862651824951,
      "learning_rate": 1.527777777777778e-06,
      "loss": 0.0629,
      "step": 70
    },
    {
      "epoch": 31.09090909090909,
      "eval_accuracy": 0.9964285714285714,
      "eval_loss": 0.08421190828084946,
      "eval_runtime": 8.3691,
      "eval_samples_per_second": 66.913,
      "eval_steps_per_second": 8.364,
      "step": 70
    },
    {
      "epoch": 35.45454545454545,
      "grad_norm": 5.019294738769531,
      "learning_rate": 1.3888888888888888e-07,
      "loss": 0.049,
      "step": 80
    },
    {
      "epoch": 35.45454545454545,
      "eval_accuracy": 0.9964285714285714,
      "eval_loss": 0.07256519049406052,
      "eval_runtime": 8.7265,
      "eval_samples_per_second": 64.172,
      "eval_steps_per_second": 8.022,
      "step": 80
    },
    {
      "epoch": 35.45454545454545,
      "step": 80,
      "total_flos": 7.012786101918106e+17,
      "train_loss": 0.4508287474513054,
      "train_runtime": 460.7942,
      "train_samples_per_second": 7.031,
      "train_steps_per_second": 0.174
    }
  ],
  "logging_steps": 10,
  "max_steps": 80,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 40,
  "save_steps": 10,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 7.012786101918106e+17,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}