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

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 6.0,
  "eval_steps": 1,
  "global_step": 10,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 156.6267852783203,
      "learning_rate": 2.5e-05,
      "loss": 0.9475,
      "step": 1
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.48,
      "eval_loss": 1.2892873287200928,
      "eval_runtime": 16.5816,
      "eval_samples_per_second": 15.077,
      "eval_steps_per_second": 0.784,
      "step": 1
    },
    {
      "epoch": 1.28,
      "grad_norm": 49.83559799194336,
      "learning_rate": 5e-05,
      "loss": 0.2955,
      "step": 2
    },
    {
      "epoch": 1.28,
      "eval_accuracy": 0.52,
      "eval_loss": 1.6241645812988281,
      "eval_runtime": 16.5757,
      "eval_samples_per_second": 15.082,
      "eval_steps_per_second": 0.784,
      "step": 2
    },
    {
      "epoch": 2.0,
      "grad_norm": 155.0616455078125,
      "learning_rate": 4.375e-05,
      "loss": 1.0042,
      "step": 3
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.516,
      "eval_loss": 2.1022417545318604,
      "eval_runtime": 16.5771,
      "eval_samples_per_second": 15.081,
      "eval_steps_per_second": 0.784,
      "step": 3
    },
    {
      "epoch": 2.56,
      "grad_norm": 126.85425567626953,
      "learning_rate": 3.7500000000000003e-05,
      "loss": 0.9628,
      "step": 4
    },
    {
      "epoch": 2.56,
      "eval_accuracy": 0.496,
      "eval_loss": 1.014886498451233,
      "eval_runtime": 16.5822,
      "eval_samples_per_second": 15.076,
      "eval_steps_per_second": 0.784,
      "step": 4
    },
    {
      "epoch": 3.0,
      "grad_norm": 59.33885955810547,
      "learning_rate": 3.125e-05,
      "loss": 0.3604,
      "step": 5
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.528,
      "eval_loss": 1.160253643989563,
      "eval_runtime": 16.5945,
      "eval_samples_per_second": 15.065,
      "eval_steps_per_second": 0.783,
      "step": 5
    },
    {
      "epoch": 3.84,
      "grad_norm": 105.5917739868164,
      "learning_rate": 2.5e-05,
      "loss": 0.7077,
      "step": 6
    },
    {
      "epoch": 3.84,
      "eval_accuracy": 0.556,
      "eval_loss": 1.1685298681259155,
      "eval_runtime": 16.5804,
      "eval_samples_per_second": 15.078,
      "eval_steps_per_second": 0.784,
      "step": 6
    },
    {
      "epoch": 4.0,
      "grad_norm": 17.550357818603516,
      "learning_rate": 1.8750000000000002e-05,
      "loss": 0.147,
      "step": 7
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.548,
      "eval_loss": 0.9757280349731445,
      "eval_runtime": 16.5787,
      "eval_samples_per_second": 15.08,
      "eval_steps_per_second": 0.784,
      "step": 7
    },
    {
      "epoch": 5.0,
      "grad_norm": 34.28722381591797,
      "learning_rate": 1.25e-05,
      "loss": 0.7256,
      "step": 8
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.548,
      "eval_loss": 0.8535332679748535,
      "eval_runtime": 16.6122,
      "eval_samples_per_second": 15.049,
      "eval_steps_per_second": 0.783,
      "step": 8
    },
    {
      "epoch": 5.12,
      "grad_norm": 1.853622555732727,
      "learning_rate": 6.25e-06,
      "loss": 0.0673,
      "step": 9
    },
    {
      "epoch": 5.12,
      "eval_accuracy": 0.576,
      "eval_loss": 0.788220226764679,
      "eval_runtime": 16.5694,
      "eval_samples_per_second": 15.088,
      "eval_steps_per_second": 0.785,
      "step": 9
    },
    {
      "epoch": 6.0,
      "grad_norm": 32.69328308105469,
      "learning_rate": 0.0,
      "loss": 0.5355,
      "step": 10
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.576,
      "eval_loss": 0.744477391242981,
      "eval_runtime": 16.5965,
      "eval_samples_per_second": 15.063,
      "eval_steps_per_second": 0.783,
      "step": 10
    },
    {
      "epoch": 6.0,
      "step": 10,
      "total_flos": 1.851279419879424e+16,
      "train_loss": 0.5753317460417747,
      "train_runtime": 1470.2381,
      "train_samples_per_second": 6.802,
      "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": 1.851279419879424e+16,
  "train_batch_size": 20,
  "trial_name": null,
  "trial_params": null
}