trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.9890909090909091,
  "eval_steps": 500,
  "global_step": 68,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.07,
      "learning_rate": 0.0001973438054360693,
      "loss": 1.1721,
      "step": 5
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00018951632913550626,
      "loss": 1.0736,
      "step": 10
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00017693339709828792,
      "loss": 1.0295,
      "step": 15
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00016026346363792567,
      "loss": 1.0088,
      "step": 20
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00014039210048718949,
      "loss": 0.9961,
      "step": 25
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00011837495178165706,
      "loss": 0.9801,
      "step": 30
    },
    {
      "epoch": 0.51,
      "learning_rate": 9.538165413542607e-05,
      "loss": 0.9603,
      "step": 35
    },
    {
      "epoch": 0.58,
      "learning_rate": 7.263370099279172e-05,
      "loss": 0.9647,
      "step": 40
    },
    {
      "epoch": 0.65,
      "learning_rate": 5.1339552143314384e-05,
      "loss": 0.9355,
      "step": 45
    },
    {
      "epoch": 0.73,
      "learning_rate": 3.263043563534428e-05,
      "loss": 0.9217,
      "step": 50
    },
    {
      "epoch": 0.8,
      "learning_rate": 1.750025254016978e-05,
      "loss": 0.9623,
      "step": 55
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.75277705956443e-06,
      "loss": 0.9369,
      "step": 60
    },
    {
      "epoch": 0.95,
      "learning_rate": 9.589569124794916e-07,
      "loss": 0.9075,
      "step": 65
    },
    {
      "epoch": 0.99,
      "step": 68,
      "total_flos": 9.5110421903573e+16,
      "train_loss": 0.9853932401713204,
      "train_runtime": 6464.5656,
      "train_samples_per_second": 0.68,
      "train_steps_per_second": 0.011
    }
  ],
  "logging_steps": 5,
  "max_steps": 68,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 100,
  "total_flos": 9.5110421903573e+16,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}