trainer_state-mistral-bnb8-QLORA-super_glue-wic-sequence_classification.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 6.4,
  "eval_steps": 1,
  "global_step": 10,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.64,
      "grad_norm": 436.4122314453125,
      "learning_rate": 2.5e-05,
      "loss": 5.8587,
      "step": 1
    },
    {
      "epoch": 0.64,
      "eval_accuracy": 0.516,
      "eval_loss": 5.887912750244141,
      "eval_runtime": 13.1544,
      "eval_samples_per_second": 19.005,
      "eval_steps_per_second": 0.988,
      "step": 1
    },
    {
      "epoch": 1.28,
      "grad_norm": 448.7404479980469,
      "learning_rate": 5e-05,
      "loss": 5.979,
      "step": 2
    },
    {
      "epoch": 1.28,
      "eval_accuracy": 0.484,
      "eval_loss": 3.829352378845215,
      "eval_runtime": 13.1532,
      "eval_samples_per_second": 19.007,
      "eval_steps_per_second": 0.988,
      "step": 2
    },
    {
      "epoch": 1.92,
      "grad_norm": 345.4527282714844,
      "learning_rate": 4.375e-05,
      "loss": 3.48,
      "step": 3
    },
    {
      "epoch": 1.92,
      "eval_accuracy": 0.616,
      "eval_loss": 0.9900831580162048,
      "eval_runtime": 13.1349,
      "eval_samples_per_second": 19.033,
      "eval_steps_per_second": 0.99,
      "step": 3
    },
    {
      "epoch": 2.56,
      "grad_norm": 34.28495407104492,
      "learning_rate": 3.7500000000000003e-05,
      "loss": 0.9253,
      "step": 4
    },
    {
      "epoch": 2.56,
      "eval_accuracy": 0.484,
      "eval_loss": 2.8530449867248535,
      "eval_runtime": 13.1214,
      "eval_samples_per_second": 19.053,
      "eval_steps_per_second": 0.991,
      "step": 4
    },
    {
      "epoch": 3.2,
      "grad_norm": 369.2248229980469,
      "learning_rate": 3.125e-05,
      "loss": 2.7248,
      "step": 5
    },
    {
      "epoch": 3.2,
      "eval_accuracy": 0.484,
      "eval_loss": 0.8809528946876526,
      "eval_runtime": 13.1143,
      "eval_samples_per_second": 19.063,
      "eval_steps_per_second": 0.991,
      "step": 5
    },
    {
      "epoch": 3.84,
      "grad_norm": 57.836917877197266,
      "learning_rate": 2.5e-05,
      "loss": 0.8653,
      "step": 6
    },
    {
      "epoch": 3.84,
      "eval_accuracy": 0.516,
      "eval_loss": 2.1503427028656006,
      "eval_runtime": 13.1024,
      "eval_samples_per_second": 19.081,
      "eval_steps_per_second": 0.992,
      "step": 6
    },
    {
      "epoch": 4.48,
      "grad_norm": 308.1085205078125,
      "learning_rate": 1.8750000000000002e-05,
      "loss": 2.14,
      "step": 7
    },
    {
      "epoch": 4.48,
      "eval_accuracy": 0.516,
      "eval_loss": 1.6527570486068726,
      "eval_runtime": 13.106,
      "eval_samples_per_second": 19.075,
      "eval_steps_per_second": 0.992,
      "step": 7
    },
    {
      "epoch": 5.12,
      "grad_norm": 301.7503356933594,
      "learning_rate": 1.25e-05,
      "loss": 1.772,
      "step": 8
    },
    {
      "epoch": 5.12,
      "eval_accuracy": 0.644,
      "eval_loss": 0.7859167456626892,
      "eval_runtime": 13.1191,
      "eval_samples_per_second": 19.056,
      "eval_steps_per_second": 0.991,
      "step": 8
    },
    {
      "epoch": 5.76,
      "grad_norm": 123.35485076904297,
      "learning_rate": 6.25e-06,
      "loss": 0.786,
      "step": 9
    },
    {
      "epoch": 5.76,
      "eval_accuracy": 0.628,
      "eval_loss": 0.9045919179916382,
      "eval_runtime": 13.1146,
      "eval_samples_per_second": 19.063,
      "eval_steps_per_second": 0.991,
      "step": 9
    },
    {
      "epoch": 6.4,
      "grad_norm": 109.03207397460938,
      "learning_rate": 0.0,
      "loss": 0.7251,
      "step": 10
    },
    {
      "epoch": 6.4,
      "eval_accuracy": 0.624,
      "eval_loss": 1.0280524492263794,
      "eval_runtime": 13.1179,
      "eval_samples_per_second": 19.058,
      "eval_steps_per_second": 0.991,
      "step": 10
    },
    {
      "epoch": 6.4,
      "step": 10,
      "total_flos": 1.974162432e+16,
      "train_loss": 2.5256052255630492,
      "train_runtime": 1346.5171,
      "train_samples_per_second": 7.427,
      "train_steps_per_second": 0.007
    }
  ],
  "logging_steps": 1,
  "max_steps": 10,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 10,
  "save_steps": 500,
  "total_flos": 1.974162432e+16,
  "train_batch_size": 20,
  "trial_name": null,
  "trial_params": null
}