trainer_state.json

{
  "best_metric": 0.8696532845497131,
  "best_model_checkpoint": "/users/zyong2/data/zyong2/bigscience/data/processed/024/bloom-350m_si_fish_100000samples_-1vocab_original-frozen/checkpoint-25000",
  "epoch": 0.4656230502034773,
  "global_step": 25000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.05,
      "learning_rate": 9e-05,
      "loss": 1.1743,
      "step": 2500
    },
    {
      "epoch": 0.09,
      "learning_rate": 8e-05,
      "loss": 1.0159,
      "step": 5000
    },
    {
      "epoch": 0.09,
      "eval_loss": 0.9803577661514282,
      "eval_runtime": 2281.7823,
      "eval_samples_per_second": 9.532,
      "eval_steps_per_second": 4.766,
      "step": 5000
    },
    {
      "epoch": 0.14,
      "learning_rate": 7e-05,
      "loss": 0.9658,
      "step": 7500
    },
    {
      "epoch": 0.19,
      "learning_rate": 6e-05,
      "loss": 0.942,
      "step": 10000
    },
    {
      "epoch": 0.19,
      "eval_loss": 0.9261226654052734,
      "eval_runtime": 2271.7319,
      "eval_samples_per_second": 9.575,
      "eval_steps_per_second": 4.788,
      "step": 10000
    },
    {
      "epoch": 0.23,
      "learning_rate": 5e-05,
      "loss": 0.9232,
      "step": 12500
    },
    {
      "epoch": 0.28,
      "learning_rate": 4e-05,
      "loss": 0.9112,
      "step": 15000
    },
    {
      "epoch": 0.28,
      "eval_loss": 0.8969926834106445,
      "eval_runtime": 2271.0001,
      "eval_samples_per_second": 9.578,
      "eval_steps_per_second": 4.789,
      "step": 15000
    },
    {
      "epoch": 0.33,
      "learning_rate": 3e-05,
      "loss": 0.8986,
      "step": 17500
    },
    {
      "epoch": 0.37,
      "learning_rate": 2e-05,
      "loss": 0.8916,
      "step": 20000
    },
    {
      "epoch": 0.37,
      "eval_loss": 0.8803833723068237,
      "eval_runtime": 2272.3302,
      "eval_samples_per_second": 9.572,
      "eval_steps_per_second": 4.786,
      "step": 20000
    },
    {
      "epoch": 0.42,
      "learning_rate": 1e-05,
      "loss": 0.8837,
      "step": 22500
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0,
      "loss": 0.8791,
      "step": 25000
    },
    {
      "epoch": 0.47,
      "eval_loss": 0.8696532845497131,
      "eval_runtime": 2270.8318,
      "eval_samples_per_second": 9.578,
      "eval_steps_per_second": 4.789,
      "step": 25000
    },
    {
      "epoch": 0.47,
      "step": 25000,
      "total_flos": 3.714827943936e+17,
      "train_loss": 0.948534130859375,
      "train_runtime": 69546.3409,
      "train_samples_per_second": 2.876,
      "train_steps_per_second": 0.359
    }
  ],
  "max_steps": 25000,
  "num_train_epochs": 1,
  "total_flos": 3.714827943936e+17,
  "trial_name": null,
  "trial_params": null
}