trainer_state.json

{
  "best_metric": 0.8214285714285714,
  "best_model_checkpoint": "videomae-base-finetuned-good-gestureUnitV11/checkpoint-34",
  "epoch": 4.15,
  "eval_steps": 500,
  "global_step": 80,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.125,
      "grad_norm": 7.015880107879639,
      "learning_rate": 4.8611111111111115e-05,
      "loss": 0.6685,
      "step": 10
    },
    {
      "epoch": 0.2125,
      "eval_accuracy": 0.6071428571428571,
      "eval_accuracy_gunit": 1.0,
      "eval_accuracy_nothing": 0.2903225806451613,
      "eval_loss": 0.6297298073768616,
      "eval_runtime": 7.5848,
      "eval_samples_per_second": 7.383,
      "eval_steps_per_second": 0.923,
      "step": 17
    },
    {
      "epoch": 1.0375,
      "grad_norm": 5.0800676345825195,
      "learning_rate": 4.166666666666667e-05,
      "loss": 0.6397,
      "step": 20
    },
    {
      "epoch": 1.1625,
      "grad_norm": 4.701914310455322,
      "learning_rate": 3.472222222222222e-05,
      "loss": 0.5348,
      "step": 30
    },
    {
      "epoch": 1.2125,
      "eval_accuracy": 0.8214285714285714,
      "eval_accuracy_gunit": 0.6,
      "eval_accuracy_nothing": 1.0,
      "eval_loss": 0.42920514941215515,
      "eval_runtime": 7.3236,
      "eval_samples_per_second": 7.647,
      "eval_steps_per_second": 0.956,
      "step": 34
    },
    {
      "epoch": 2.075,
      "grad_norm": 26.41465187072754,
      "learning_rate": 2.777777777777778e-05,
      "loss": 0.4109,
      "step": 40
    },
    {
      "epoch": 2.2,
      "grad_norm": 12.76979923248291,
      "learning_rate": 2.0833333333333336e-05,
      "loss": 0.3395,
      "step": 50
    },
    {
      "epoch": 2.2125,
      "eval_accuracy": 0.7857142857142857,
      "eval_accuracy_gunit": 0.88,
      "eval_accuracy_nothing": 0.7096774193548387,
      "eval_loss": 0.4852010905742645,
      "eval_runtime": 6.9927,
      "eval_samples_per_second": 8.008,
      "eval_steps_per_second": 1.001,
      "step": 51
    },
    {
      "epoch": 3.1125,
      "grad_norm": 15.371440887451172,
      "learning_rate": 1.388888888888889e-05,
      "loss": 0.3607,
      "step": 60
    },
    {
      "epoch": 3.2125,
      "eval_accuracy": 0.8214285714285714,
      "eval_accuracy_gunit": 0.6,
      "eval_accuracy_nothing": 1.0,
      "eval_loss": 0.4506545662879944,
      "eval_runtime": 6.8621,
      "eval_samples_per_second": 8.161,
      "eval_steps_per_second": 1.02,
      "step": 68
    },
    {
      "epoch": 4.025,
      "grad_norm": 11.717324256896973,
      "learning_rate": 6.944444444444445e-06,
      "loss": 0.3328,
      "step": 70
    },
    {
      "epoch": 4.15,
      "grad_norm": 13.87458610534668,
      "learning_rate": 0.0,
      "loss": 0.2436,
      "step": 80
    },
    {
      "epoch": 4.15,
      "eval_accuracy": 0.7857142857142857,
      "eval_accuracy_gunit": 0.76,
      "eval_accuracy_nothing": 0.8064516129032258,
      "eval_loss": 0.44327452778816223,
      "eval_runtime": 6.9163,
      "eval_samples_per_second": 8.097,
      "eval_steps_per_second": 1.012,
      "step": 80
    },
    {
      "epoch": 4.15,
      "step": 80,
      "total_flos": 7.924970789040292e+17,
      "train_loss": 0.441299644112587,
      "train_runtime": 198.5865,
      "train_samples_per_second": 3.223,
      "train_steps_per_second": 0.403
    },
    {
      "epoch": 4.15,
      "eval_accuracy": 0.9310344827586207,
      "eval_accuracy_gunit": 0.8571428571428571,
      "eval_accuracy_nothing": 1.0,
      "eval_loss": 0.30250757932662964,
      "eval_runtime": 12.2177,
      "eval_samples_per_second": 7.121,
      "eval_steps_per_second": 0.9,
      "step": 80
    },
    {
      "epoch": 4.15,
      "eval_accuracy": 0.9310344827586207,
      "eval_accuracy_gunit": 0.8571428571428571,
      "eval_accuracy_nothing": 1.0,
      "eval_loss": 0.3025076389312744,
      "eval_runtime": 10.959,
      "eval_samples_per_second": 7.939,
      "eval_steps_per_second": 1.004,
      "step": 80
    }
  ],
  "logging_steps": 10,
  "max_steps": 80,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 9223372036854775807,
  "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": 7.924970789040292e+17,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}