run-27/checkpoint-240/trainer_state.json

{
  "best_metric": 0.8277886497064579,
  "best_model_checkpoint": "tiny-bert-sst2-distilled/run-27/checkpoint-240",
  "epoch": 5.0,
  "eval_steps": 500,
  "global_step": 240,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 3.0442044734954834,
      "learning_rate": 0.00047720243978875736,
      "loss": 0.5494,
      "step": 48
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.7270058708414873,
      "eval_f1": 0.7787470261697065,
      "eval_loss": 0.4914495050907135,
      "eval_precision": 0.6546666666666666,
      "eval_recall": 0.9608610567514677,
      "eval_runtime": 28.018,
      "eval_samples_per_second": 36.476,
      "eval_steps_per_second": 0.571,
      "step": 48
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.9368475675582886,
      "learning_rate": 0.000357901829841568,
      "loss": 0.4641,
      "step": 96
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.7720156555772995,
      "eval_f1": 0.8059950041631974,
      "eval_loss": 0.4658460021018982,
      "eval_precision": 0.7014492753623188,
      "eval_recall": 0.9471624266144814,
      "eval_runtime": 27.3749,
      "eval_samples_per_second": 37.333,
      "eval_steps_per_second": 0.584,
      "step": 96
    },
    {
      "epoch": 3.0,
      "grad_norm": 3.3749494552612305,
      "learning_rate": 0.00023860121989437868,
      "loss": 0.4456,
      "step": 144
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8189823874755382,
      "eval_f1": 0.8262910798122065,
      "eval_loss": 0.4206434190273285,
      "eval_precision": 0.7942238267148014,
      "eval_recall": 0.8610567514677103,
      "eval_runtime": 27.471,
      "eval_samples_per_second": 37.203,
      "eval_steps_per_second": 0.582,
      "step": 144
    },
    {
      "epoch": 4.0,
      "grad_norm": 3.301919937133789,
      "learning_rate": 0.00011930060994718934,
      "loss": 0.4162,
      "step": 192
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8228962818003914,
      "eval_f1": 0.8281101614434948,
      "eval_loss": 0.4116476774215698,
      "eval_precision": 0.8044280442804428,
      "eval_recall": 0.8532289628180039,
      "eval_runtime": 27.5336,
      "eval_samples_per_second": 37.118,
      "eval_steps_per_second": 0.581,
      "step": 192
    },
    {
      "epoch": 5.0,
      "grad_norm": 1.5170879364013672,
      "learning_rate": 0.0,
      "loss": 0.3983,
      "step": 240
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.8277886497064579,
      "eval_f1": 0.8376383763837638,
      "eval_loss": 0.4083719253540039,
      "eval_precision": 0.7923211169284468,
      "eval_recall": 0.8884540117416829,
      "eval_runtime": 27.5311,
      "eval_samples_per_second": 37.122,
      "eval_steps_per_second": 0.581,
      "step": 240
    }
  ],
  "logging_steps": 500,
  "max_steps": 240,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 5,
  "save_steps": 500,
  "total_flos": 1178475986400.0,
  "train_batch_size": 64,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.8647030881639047,
    "learning_rate": 0.0005965030497359467,
    "num_train_epochs": 5,
    "temperature": 29
  }
}