trainer_state.json

{
  "best_metric": 0.00012142229388700798,
  "best_model_checkpoint": "data/further_pretraining/roberta-large/checkpoint-9356",
  "epoch": 2.0,
  "global_step": 9356,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.11,
      "learning_rate": 1.9786233433091067e-05,
      "loss": 0.0753,
      "step": 500
    },
    {
      "epoch": 0.21,
      "learning_rate": 1.9572466866182133e-05,
      "loss": 0.0589,
      "step": 1000
    },
    {
      "epoch": 0.32,
      "learning_rate": 1.9358700299273195e-05,
      "loss": 0.0558,
      "step": 1500
    },
    {
      "epoch": 0.43,
      "learning_rate": 1.914493373236426e-05,
      "loss": 0.0551,
      "step": 2000
    },
    {
      "epoch": 0.53,
      "learning_rate": 1.8931167165455325e-05,
      "loss": 0.0541,
      "step": 2500
    },
    {
      "epoch": 0.64,
      "learning_rate": 1.8717400598546387e-05,
      "loss": 0.0385,
      "step": 3000
    },
    {
      "epoch": 0.75,
      "learning_rate": 1.8503634031637453e-05,
      "loss": 0.0123,
      "step": 3500
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.828986746472852e-05,
      "loss": 0.0037,
      "step": 4000
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.807610089781958e-05,
      "loss": 0.0029,
      "step": 4500
    },
    {
      "epoch": 1.0,
      "eval_loss": 0.0009945582132786512,
      "eval_runtime": 755.7373,
      "eval_samples_per_second": 99.032,
      "eval_steps_per_second": 6.19,
      "step": 4678
    },
    {
      "epoch": 1.07,
      "learning_rate": 1.7862334330910646e-05,
      "loss": 0.0017,
      "step": 5000
    },
    {
      "epoch": 1.18,
      "learning_rate": 1.764856776400171e-05,
      "loss": 0.0011,
      "step": 5500
    },
    {
      "epoch": 1.28,
      "learning_rate": 1.7434801197092777e-05,
      "loss": 0.0136,
      "step": 6000
    },
    {
      "epoch": 1.39,
      "learning_rate": 1.7221034630183842e-05,
      "loss": 0.001,
      "step": 6500
    },
    {
      "epoch": 1.5,
      "learning_rate": 1.7007268063274904e-05,
      "loss": 0.0009,
      "step": 7000
    },
    {
      "epoch": 1.6,
      "learning_rate": 1.679350149636597e-05,
      "loss": 0.0007,
      "step": 7500
    },
    {
      "epoch": 1.71,
      "learning_rate": 1.6579734929457035e-05,
      "loss": 0.0007,
      "step": 8000
    },
    {
      "epoch": 1.82,
      "learning_rate": 1.63659683625481e-05,
      "loss": 0.0013,
      "step": 8500
    },
    {
      "epoch": 1.92,
      "learning_rate": 1.6152201795639163e-05,
      "loss": 0.0012,
      "step": 9000
    },
    {
      "epoch": 2.0,
      "eval_loss": 0.00012142229388700798,
      "eval_runtime": 758.5152,
      "eval_samples_per_second": 98.669,
      "eval_steps_per_second": 6.167,
      "step": 9356
    }
  ],
  "max_steps": 46780,
  "num_train_epochs": 10,
  "total_flos": 3.4879581646470144e+16,
  "trial_name": null,
  "trial_params": null
}