{
  "best_metric": 0.8238747553816047,
  "best_model_checkpoint": "tiny-bert-sst2-distilled/run-4/checkpoint-768",
  "epoch": 9.0,
  "eval_steps": 500,
  "global_step": 864,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 1.90240478515625,
      "learning_rate": 0.0008040462940429662,
      "loss": 0.4563,
      "step": 96
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.7446183953033269,
      "eval_f1": 0.7473378509196515,
      "eval_loss": 0.415781170129776,
      "eval_precision": 0.7394636015325671,
      "eval_recall": 0.7553816046966731,
      "eval_runtime": 25.3619,
      "eval_samples_per_second": 40.297,
      "eval_steps_per_second": 1.262,
      "step": 96
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.7548394799232483,
      "learning_rate": 0.0007147078169270811,
      "loss": 0.4243,
      "step": 192
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.7896281800391389,
      "eval_f1": 0.7969782813975448,
      "eval_loss": 0.39640527963638306,
      "eval_precision": 0.7700729927007299,
      "eval_recall": 0.8258317025440313,
      "eval_runtime": 25.8877,
      "eval_samples_per_second": 39.478,
      "eval_steps_per_second": 1.236,
      "step": 192
    },
    {
      "epoch": 3.0,
      "grad_norm": 2.091404914855957,
      "learning_rate": 0.0006253693398111959,
      "loss": 0.4007,
      "step": 288
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8052837573385518,
      "eval_f1": 0.8240495137046862,
      "eval_loss": 0.39981088042259216,
      "eval_precision": 0.7516129032258064,
      "eval_recall": 0.9119373776908023,
      "eval_runtime": 25.886,
      "eval_samples_per_second": 39.481,
      "eval_steps_per_second": 1.236,
      "step": 288
    },
    {
      "epoch": 4.0,
      "grad_norm": 1.994834542274475,
      "learning_rate": 0.0005360308626953108,
      "loss": 0.3871,
      "step": 384
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8082191780821918,
      "eval_f1": 0.8175046554934823,
      "eval_loss": 0.38285842537879944,
      "eval_precision": 0.7797513321492007,
      "eval_recall": 0.8590998043052838,
      "eval_runtime": 25.8573,
      "eval_samples_per_second": 39.525,
      "eval_steps_per_second": 1.238,
      "step": 384
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.7219040393829346,
      "learning_rate": 0.00044669238557942565,
      "loss": 0.378,
      "step": 480
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.773972602739726,
      "eval_f1": 0.7640449438202248,
      "eval_loss": 0.39512935280799866,
      "eval_precision": 0.7991452991452992,
      "eval_recall": 0.7318982387475538,
      "eval_runtime": 25.8711,
      "eval_samples_per_second": 39.503,
      "eval_steps_per_second": 1.237,
      "step": 480
    },
    {
      "epoch": 6.0,
      "grad_norm": 1.2725290060043335,
      "learning_rate": 0.00035735390846354054,
      "loss": 0.3759,
      "step": 576
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.815068493150685,
      "eval_f1": 0.8270814272644098,
      "eval_loss": 0.3832918703556061,
      "eval_precision": 0.7766323024054983,
      "eval_recall": 0.8845401174168297,
      "eval_runtime": 26.2709,
      "eval_samples_per_second": 38.902,
      "eval_steps_per_second": 1.218,
      "step": 576
    },
    {
      "epoch": 7.0,
      "grad_norm": 0.7257367968559265,
      "learning_rate": 0.0002680154313476554,
      "loss": 0.3679,
      "step": 672
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.815068493150685,
      "eval_f1": 0.8280254777070064,
      "eval_loss": 0.37927353382110596,
      "eval_precision": 0.7738095238095238,
      "eval_recall": 0.8904109589041096,
      "eval_runtime": 25.5612,
      "eval_samples_per_second": 39.982,
      "eval_steps_per_second": 1.252,
      "step": 672
    },
    {
      "epoch": 8.0,
      "grad_norm": 1.1571022272109985,
      "learning_rate": 0.00017867695423177027,
      "loss": 0.3609,
      "step": 768
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.8238747553816047,
      "eval_f1": 0.832089552238806,
      "eval_loss": 0.37364885210990906,
      "eval_precision": 0.7950089126559715,
      "eval_recall": 0.87279843444227,
      "eval_runtime": 25.5421,
      "eval_samples_per_second": 40.012,
      "eval_steps_per_second": 1.253,
      "step": 768
    },
    {
      "epoch": 9.0,
      "grad_norm": 0.6674935817718506,
      "learning_rate": 8.933847711588513e-05,
      "loss": 0.3556,
      "step": 864
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.8238747553816047,
      "eval_f1": 0.8314606741573034,
      "eval_loss": 0.3790241479873657,
      "eval_precision": 0.7971274685816876,
      "eval_recall": 0.8688845401174168,
      "eval_runtime": 26.0614,
      "eval_samples_per_second": 39.215,
      "eval_steps_per_second": 1.228,
      "step": 864
    }
  ],
  "logging_steps": 500,
  "max_steps": 960,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 10,
  "save_steps": 500,
  "total_flos": 2121256775520.0,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.6601531096042508,
    "learning_rate": 0.0008933847711588513,
    "num_train_epochs": 10,
    "temperature": 4
  }
}