trainer_state.json

{
  "best_metric": 1.0858668088912964,
  "best_model_checkpoint": "/home/khalid/Documents/github_rep/bigscience/data/processed/50/bloom-560m_my_bitfit_100000samples_-1vocab_original-frozen/checkpoint-25000",
  "epoch": 0.39631741854686253,
  "global_step": 25000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.04,
      "learning_rate": 9e-05,
      "loss": 1.3263,
      "step": 2500
    },
    {
      "epoch": 0.08,
      "learning_rate": 8e-05,
      "loss": 1.1812,
      "step": 5000
    },
    {
      "epoch": 0.08,
      "eval_loss": 1.1644270420074463,
      "eval_runtime": 2656.1309,
      "eval_samples_per_second": 9.443,
      "eval_steps_per_second": 4.722,
      "step": 5000
    },
    {
      "epoch": 0.12,
      "learning_rate": 7e-05,
      "loss": 1.1436,
      "step": 7500
    },
    {
      "epoch": 0.16,
      "learning_rate": 6e-05,
      "loss": 1.1212,
      "step": 10000
    },
    {
      "epoch": 0.16,
      "eval_loss": 1.1189237833023071,
      "eval_runtime": 2658.6123,
      "eval_samples_per_second": 9.434,
      "eval_steps_per_second": 4.717,
      "step": 10000
    },
    {
      "epoch": 0.2,
      "learning_rate": 5e-05,
      "loss": 1.1092,
      "step": 12500
    },
    {
      "epoch": 0.24,
      "learning_rate": 4e-05,
      "loss": 1.0977,
      "step": 15000
    },
    {
      "epoch": 0.24,
      "eval_loss": 1.0991963148117065,
      "eval_runtime": 2656.572,
      "eval_samples_per_second": 9.441,
      "eval_steps_per_second": 4.721,
      "step": 15000
    },
    {
      "epoch": 0.28,
      "learning_rate": 3e-05,
      "loss": 1.0908,
      "step": 17500
    },
    {
      "epoch": 0.32,
      "learning_rate": 2e-05,
      "loss": 1.0856,
      "step": 20000
    },
    {
      "epoch": 0.32,
      "eval_loss": 1.0892250537872314,
      "eval_runtime": 2657.3128,
      "eval_samples_per_second": 9.438,
      "eval_steps_per_second": 4.719,
      "step": 20000
    },
    {
      "epoch": 0.36,
      "learning_rate": 1e-05,
      "loss": 1.0825,
      "step": 22500
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0,
      "loss": 1.08,
      "step": 25000
    },
    {
      "epoch": 0.4,
      "eval_loss": 1.0858668088912964,
      "eval_runtime": 2656.2605,
      "eval_samples_per_second": 9.442,
      "eval_steps_per_second": 4.721,
      "step": 25000
    },
    {
      "epoch": 0.4,
      "step": 25000,
      "total_flos": 3.714827943936e+17,
      "train_loss": 1.1317986328125,
      "train_runtime": 62833.7719,
      "train_samples_per_second": 3.183,
      "train_steps_per_second": 0.398
    }
  ],
  "max_steps": 25000,
  "num_train_epochs": 1,
  "total_flos": 3.714827943936e+17,
  "trial_name": null,
  "trial_params": null
}