trainer_state.json

{
  "best_metric": 0.7268518518518519,
  "best_model_checkpoint": "swinv2-large-patch4-window12to24-192to384-22kto1k-ft-microbes-merged\\checkpoint-106",
  "epoch": 7.868852459016393,
  "eval_steps": 500,
  "global_step": 120,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.66,
      "learning_rate": 4.166666666666667e-05,
      "loss": 3.8355,
      "step": 10
    },
    {
      "epoch": 0.98,
      "eval_accuracy": 0.3333333333333333,
      "eval_loss": 2.583052158355713,
      "eval_runtime": 271.6063,
      "eval_samples_per_second": 0.795,
      "eval_steps_per_second": 0.026,
      "step": 15
    },
    {
      "epoch": 1.31,
      "learning_rate": 4.62962962962963e-05,
      "loss": 2.7166,
      "step": 20
    },
    {
      "epoch": 1.97,
      "learning_rate": 4.166666666666667e-05,
      "loss": 1.9292,
      "step": 30
    },
    {
      "epoch": 1.97,
      "eval_accuracy": 0.5046296296296297,
      "eval_loss": 1.6849619150161743,
      "eval_runtime": 268.675,
      "eval_samples_per_second": 0.804,
      "eval_steps_per_second": 0.026,
      "step": 30
    },
    {
      "epoch": 2.62,
      "learning_rate": 3.7037037037037037e-05,
      "loss": 1.4121,
      "step": 40
    },
    {
      "epoch": 2.95,
      "eval_accuracy": 0.5972222222222222,
      "eval_loss": 1.2323545217514038,
      "eval_runtime": 260.3765,
      "eval_samples_per_second": 0.83,
      "eval_steps_per_second": 0.027,
      "step": 45
    },
    {
      "epoch": 3.28,
      "learning_rate": 3.240740740740741e-05,
      "loss": 1.1566,
      "step": 50
    },
    {
      "epoch": 3.93,
      "learning_rate": 2.777777777777778e-05,
      "loss": 1.0121,
      "step": 60
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.6851851851851852,
      "eval_loss": 1.0345041751861572,
      "eval_runtime": 260.8934,
      "eval_samples_per_second": 0.828,
      "eval_steps_per_second": 0.027,
      "step": 61
    },
    {
      "epoch": 4.59,
      "learning_rate": 2.314814814814815e-05,
      "loss": 0.854,
      "step": 70
    },
    {
      "epoch": 4.98,
      "eval_accuracy": 0.6805555555555556,
      "eval_loss": 0.9663412570953369,
      "eval_runtime": 260.7389,
      "eval_samples_per_second": 0.828,
      "eval_steps_per_second": 0.027,
      "step": 76
    },
    {
      "epoch": 5.25,
      "learning_rate": 1.8518518518518518e-05,
      "loss": 0.8067,
      "step": 80
    },
    {
      "epoch": 5.9,
      "learning_rate": 1.388888888888889e-05,
      "loss": 0.701,
      "step": 90
    },
    {
      "epoch": 5.97,
      "eval_accuracy": 0.6990740740740741,
      "eval_loss": 0.9587138295173645,
      "eval_runtime": 260.4872,
      "eval_samples_per_second": 0.829,
      "eval_steps_per_second": 0.027,
      "step": 91
    },
    {
      "epoch": 6.56,
      "learning_rate": 9.259259259259259e-06,
      "loss": 0.5956,
      "step": 100
    },
    {
      "epoch": 6.95,
      "eval_accuracy": 0.7268518518518519,
      "eval_loss": 0.8626322150230408,
      "eval_runtime": 261.3301,
      "eval_samples_per_second": 0.827,
      "eval_steps_per_second": 0.027,
      "step": 106
    },
    {
      "epoch": 7.21,
      "learning_rate": 4.6296296296296296e-06,
      "loss": 0.6181,
      "step": 110
    },
    {
      "epoch": 7.87,
      "learning_rate": 0.0,
      "loss": 0.5713,
      "step": 120
    },
    {
      "epoch": 7.87,
      "eval_accuracy": 0.7222222222222222,
      "eval_loss": 0.864508867263794,
      "eval_runtime": 260.0671,
      "eval_samples_per_second": 0.831,
      "eval_steps_per_second": 0.027,
      "step": 120
    },
    {
      "epoch": 7.87,
      "step": 120,
      "total_flos": 7.925444841409413e+18,
      "train_loss": 1.3507493138313293,
      "train_runtime": 51454.7075,
      "train_samples_per_second": 0.302,
      "train_steps_per_second": 0.002
    }
  ],
  "logging_steps": 10,
  "max_steps": 120,
  "num_train_epochs": 8,
  "save_steps": 500,
  "total_flos": 7.925444841409413e+18,
  "trial_name": null,
  "trial_params": null
}