trainer_state.json

{
  "best_metric": 0.8827102803738318,
  "best_model_checkpoint": "MoE-leaf-disease-convnextv2-base-22k-224/checkpoint-5",
  "epoch": 12.8,
  "eval_steps": 500,
  "global_step": 32,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.8,
      "eval_accuracy": 0.8808411214953271,
      "eval_loss": 0.46136829257011414,
      "eval_runtime": 39.2263,
      "eval_samples_per_second": 54.555,
      "eval_steps_per_second": 0.051,
      "step": 2
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8827102803738318,
      "eval_loss": 0.34400519728660583,
      "eval_runtime": 40.1116,
      "eval_samples_per_second": 53.351,
      "eval_steps_per_second": 0.05,
      "step": 5
    },
    {
      "epoch": 2.8,
      "eval_accuracy": 0.8827102803738318,
      "eval_loss": 0.344054251909256,
      "eval_runtime": 37.365,
      "eval_samples_per_second": 57.273,
      "eval_steps_per_second": 0.054,
      "step": 7
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.02522450126707554,
      "learning_rate": 3.928571428571429e-05,
      "loss": 0.3155,
      "step": 10
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8827102803738318,
      "eval_loss": 0.3439185917377472,
      "eval_runtime": 38.4322,
      "eval_samples_per_second": 55.682,
      "eval_steps_per_second": 0.052,
      "step": 10
    },
    {
      "epoch": 4.8,
      "eval_accuracy": 0.8827102803738318,
      "eval_loss": 0.34368354082107544,
      "eval_runtime": 38.0602,
      "eval_samples_per_second": 56.227,
      "eval_steps_per_second": 0.053,
      "step": 12
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.8827102803738318,
      "eval_loss": 0.343065083026886,
      "eval_runtime": 38.3046,
      "eval_samples_per_second": 55.868,
      "eval_steps_per_second": 0.052,
      "step": 15
    },
    {
      "epoch": 6.8,
      "eval_accuracy": 0.8827102803738318,
      "eval_loss": 0.3426330089569092,
      "eval_runtime": 38.0517,
      "eval_samples_per_second": 56.239,
      "eval_steps_per_second": 0.053,
      "step": 17
    },
    {
      "epoch": 8.0,
      "grad_norm": 0.007871972396969795,
      "learning_rate": 2.1428571428571428e-05,
      "loss": 0.2577,
      "step": 20
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.8827102803738318,
      "eval_loss": 0.3421165347099304,
      "eval_runtime": 38.0514,
      "eval_samples_per_second": 56.24,
      "eval_steps_per_second": 0.053,
      "step": 20
    },
    {
      "epoch": 8.8,
      "eval_accuracy": 0.8827102803738318,
      "eval_loss": 0.3418828845024109,
      "eval_runtime": 38.9753,
      "eval_samples_per_second": 54.907,
      "eval_steps_per_second": 0.051,
      "step": 22
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.8827102803738318,
      "eval_loss": 0.3416748642921448,
      "eval_runtime": 38.7299,
      "eval_samples_per_second": 55.254,
      "eval_steps_per_second": 0.052,
      "step": 25
    },
    {
      "epoch": 10.8,
      "eval_accuracy": 0.8827102803738318,
      "eval_loss": 0.3415946364402771,
      "eval_runtime": 37.8063,
      "eval_samples_per_second": 56.604,
      "eval_steps_per_second": 0.053,
      "step": 27
    },
    {
      "epoch": 12.0,
      "grad_norm": 0.01667422242462635,
      "learning_rate": 3.5714285714285714e-06,
      "loss": 0.2601,
      "step": 30
    },
    {
      "epoch": 12.0,
      "eval_accuracy": 0.8827102803738318,
      "eval_loss": 0.3415359854698181,
      "eval_runtime": 38.519,
      "eval_samples_per_second": 55.557,
      "eval_steps_per_second": 0.052,
      "step": 30
    },
    {
      "epoch": 12.8,
      "eval_accuracy": 0.8827102803738318,
      "eval_loss": 0.34151995182037354,
      "eval_runtime": 37.7802,
      "eval_samples_per_second": 56.643,
      "eval_steps_per_second": 0.053,
      "step": 32
    },
    {
      "epoch": 12.8,
      "step": 32,
      "total_flos": 0.0,
      "train_loss": 0.2766976151615381,
      "train_runtime": 2412.5227,
      "train_samples_per_second": 127.714,
      "train_steps_per_second": 0.013
    }
  ],
  "logging_steps": 10,
  "max_steps": 32,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 16,
  "save_steps": 500,
  "total_flos": 0.0,
  "train_batch_size": 2000,
  "trial_name": null,
  "trial_params": null
}