run-4/checkpoint-960/trainer_state.json

{
  "best_metric": 0.6674509803921569,
  "best_model_checkpoint": "tiny-bert-sst2-distilled/run-4/checkpoint-160",
  "epoch": 6.0,
  "eval_steps": 500,
  "global_step": 960,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 0.30153390765190125,
      "learning_rate": 0.0001806546260556293,
      "loss": 0.1164,
      "step": 160
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.6674509803921569,
      "eval_f1": 0.03636363636363636,
      "eval_loss": 0.10561967641115189,
      "eval_mcc": 0.04629100498862757,
      "eval_precision": 0.5333333333333333,
      "eval_recall": 0.018823529411764704,
      "eval_runtime": 1.857,
      "eval_samples_per_second": 686.581,
      "eval_steps_per_second": 21.54,
      "step": 160
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.2800346910953522,
      "learning_rate": 0.00014452370084450344,
      "loss": 0.1067,
      "step": 320
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.6658823529411765,
      "eval_f1": 0.049107142857142856,
      "eval_loss": 0.10315733402967453,
      "eval_mcc": 0.04166956048662825,
      "eval_precision": 0.4782608695652174,
      "eval_recall": 0.02588235294117647,
      "eval_runtime": 1.8766,
      "eval_samples_per_second": 679.417,
      "eval_steps_per_second": 21.315,
      "step": 320
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.29303914308547974,
      "learning_rate": 0.00010839277563337758,
      "loss": 0.1049,
      "step": 480
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.6603921568627451,
      "eval_f1": 0.06881720430107525,
      "eval_loss": 0.10350044816732407,
      "eval_mcc": 0.02545139051903111,
      "eval_precision": 0.4,
      "eval_recall": 0.03764705882352941,
      "eval_runtime": 1.8927,
      "eval_samples_per_second": 673.655,
      "eval_steps_per_second": 21.134,
      "step": 480
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.34386146068573,
      "learning_rate": 7.226185042225172e-05,
      "loss": 0.1043,
      "step": 640
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.6666666666666666,
      "eval_f1": 0.0534521158129176,
      "eval_loss": 0.1024671271443367,
      "eval_mcc": 0.04897021068743918,
      "eval_precision": 0.5,
      "eval_recall": 0.02823529411764706,
      "eval_runtime": 1.8897,
      "eval_samples_per_second": 674.694,
      "eval_steps_per_second": 21.167,
      "step": 640
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.266721248626709,
      "learning_rate": 3.613092521112586e-05,
      "loss": 0.1039,
      "step": 800
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.6643137254901961,
      "eval_f1": 0.06550218340611354,
      "eval_loss": 0.10257755219936371,
      "eval_mcc": 0.04191297423079544,
      "eval_precision": 0.45454545454545453,
      "eval_recall": 0.03529411764705882,
      "eval_runtime": 1.872,
      "eval_samples_per_second": 681.081,
      "eval_steps_per_second": 21.367,
      "step": 800
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.17371755838394165,
      "learning_rate": 0.0,
      "loss": 0.1031,
      "step": 960
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.6666666666666666,
      "eval_f1": 0.06593406593406594,
      "eval_loss": 0.10186242312192917,
      "eval_mcc": 0.05488212999484517,
      "eval_precision": 0.5,
      "eval_recall": 0.03529411764705882,
      "eval_runtime": 1.8824,
      "eval_samples_per_second": 677.342,
      "eval_steps_per_second": 21.25,
      "step": 960
    }
  ],
  "logging_steps": 500,
  "max_steps": 960,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 6,
  "save_steps": 500,
  "total_flos": 1750532627520.0,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.1455909339678303,
    "learning_rate": 0.00021678555126675516,
    "num_train_epochs": 6,
    "temperature": 19
  }
}