checkpoint-12/trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.96,
  "eval_steps": 3,
  "global_step": 12,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.08,
      "grad_norm": 25.49334716796875,
      "learning_rate": 2e-05,
      "loss": 6.6468,
      "step": 1
    },
    {
      "epoch": 0.08,
      "eval_loss": 7.297876834869385,
      "eval_runtime": 19.9721,
      "eval_samples_per_second": 0.601,
      "eval_steps_per_second": 0.3,
      "step": 1
    },
    {
      "epoch": 0.16,
      "grad_norm": 25.83477020263672,
      "learning_rate": 4e-05,
      "loss": 7.0344,
      "step": 2
    },
    {
      "epoch": 0.24,
      "grad_norm": 29.465993881225586,
      "learning_rate": 6e-05,
      "loss": 7.0949,
      "step": 3
    },
    {
      "epoch": 0.24,
      "eval_loss": 5.278209209442139,
      "eval_runtime": 20.1162,
      "eval_samples_per_second": 0.597,
      "eval_steps_per_second": 0.298,
      "step": 3
    },
    {
      "epoch": 0.32,
      "grad_norm": 22.357271194458008,
      "learning_rate": 8e-05,
      "loss": 5.4423,
      "step": 4
    },
    {
      "epoch": 0.4,
      "grad_norm": 20.18695068359375,
      "learning_rate": 0.0001,
      "loss": 3.2963,
      "step": 5
    },
    {
      "epoch": 0.48,
      "grad_norm": 11.463640213012695,
      "learning_rate": 0.00012,
      "loss": 1.3638,
      "step": 6
    },
    {
      "epoch": 0.48,
      "eval_loss": 0.8664708733558655,
      "eval_runtime": 20.0662,
      "eval_samples_per_second": 0.598,
      "eval_steps_per_second": 0.299,
      "step": 6
    },
    {
      "epoch": 0.56,
      "grad_norm": 5.263468265533447,
      "learning_rate": 0.00014,
      "loss": 0.6614,
      "step": 7
    },
    {
      "epoch": 0.64,
      "grad_norm": 4.55341911315918,
      "learning_rate": 0.00016,
      "loss": 0.5933,
      "step": 8
    },
    {
      "epoch": 0.72,
      "grad_norm": 6.804744720458984,
      "learning_rate": 0.00018,
      "loss": 0.53,
      "step": 9
    },
    {
      "epoch": 0.72,
      "eval_loss": 0.6486384868621826,
      "eval_runtime": 19.9911,
      "eval_samples_per_second": 0.6,
      "eval_steps_per_second": 0.3,
      "step": 9
    },
    {
      "epoch": 0.8,
      "grad_norm": 14.322916030883789,
      "learning_rate": 0.0002,
      "loss": 1.0858,
      "step": 10
    },
    {
      "epoch": 0.88,
      "grad_norm": 5.152796745300293,
      "learning_rate": 0.0001,
      "loss": 0.5521,
      "step": 11
    },
    {
      "epoch": 0.96,
      "grad_norm": 3.4037420749664307,
      "learning_rate": 0.0,
      "loss": 0.4866,
      "step": 12
    },
    {
      "epoch": 0.96,
      "eval_loss": 0.4921029508113861,
      "eval_runtime": 19.9358,
      "eval_samples_per_second": 0.602,
      "eval_steps_per_second": 0.301,
      "step": 12
    }
  ],
  "logging_steps": 1,
  "max_steps": 12,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 500,
  "total_flos": 3.394812802970419e+16,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}