checkpoint-200/trainer_state.json

{
  "best_metric": 0.7148585915565491,
  "best_model_checkpoint": "../../models/lora/llama-2-7b-multitask_cot_reflection/checkpoint-200",
  "epoch": 12.5,
  "global_step": 200,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.62,
      "learning_rate": 2.9999999999999997e-05,
      "loss": 2.4102,
      "step": 10
    },
    {
      "epoch": 1.25,
      "learning_rate": 5.9999999999999995e-05,
      "loss": 2.3324,
      "step": 20
    },
    {
      "epoch": 1.88,
      "learning_rate": 8.999999999999999e-05,
      "loss": 1.9864,
      "step": 30
    },
    {
      "epoch": 2.5,
      "learning_rate": 0.00011999999999999999,
      "loss": 1.6398,
      "step": 40
    },
    {
      "epoch": 3.12,
      "learning_rate": 0.00015,
      "loss": 1.2703,
      "step": 50
    },
    {
      "epoch": 3.75,
      "learning_rate": 0.00017999999999999998,
      "loss": 1.0435,
      "step": 60
    },
    {
      "epoch": 4.38,
      "learning_rate": 0.00020999999999999998,
      "loss": 0.9847,
      "step": 70
    },
    {
      "epoch": 5.0,
      "learning_rate": 0.00023999999999999998,
      "loss": 0.956,
      "step": 80
    },
    {
      "epoch": 5.62,
      "learning_rate": 0.00027,
      "loss": 0.9321,
      "step": 90
    },
    {
      "epoch": 6.25,
      "learning_rate": 0.0003,
      "loss": 0.8775,
      "step": 100
    },
    {
      "epoch": 6.88,
      "learning_rate": 0.0002921052631578947,
      "loss": 0.818,
      "step": 110
    },
    {
      "epoch": 7.5,
      "learning_rate": 0.0002842105263157894,
      "loss": 0.8128,
      "step": 120
    },
    {
      "epoch": 8.12,
      "learning_rate": 0.0002763157894736842,
      "loss": 0.78,
      "step": 130
    },
    {
      "epoch": 8.75,
      "learning_rate": 0.0002684210526315789,
      "loss": 0.7734,
      "step": 140
    },
    {
      "epoch": 9.38,
      "learning_rate": 0.0002605263157894737,
      "loss": 0.7761,
      "step": 150
    },
    {
      "epoch": 10.0,
      "learning_rate": 0.00025263157894736836,
      "loss": 0.7668,
      "step": 160
    },
    {
      "epoch": 10.62,
      "learning_rate": 0.00024473684210526314,
      "loss": 0.7732,
      "step": 170
    },
    {
      "epoch": 11.25,
      "learning_rate": 0.00023684210526315788,
      "loss": 0.7536,
      "step": 180
    },
    {
      "epoch": 11.88,
      "learning_rate": 0.00022894736842105263,
      "loss": 0.7406,
      "step": 190
    },
    {
      "epoch": 12.5,
      "learning_rate": 0.00022105263157894733,
      "loss": 0.7288,
      "step": 200
    },
    {
      "epoch": 12.5,
      "eval_loss": 0.7148585915565491,
      "eval_runtime": 1.0356,
      "eval_samples_per_second": 20.277,
      "eval_steps_per_second": 2.897,
      "step": 200
    }
  ],
  "max_steps": 480,
  "num_train_epochs": 30,
  "total_flos": 2.6708116680081408e+17,
  "trial_name": null,
  "trial_params": null
}