trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 4.0,
  "global_step": 7348,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.27,
      "learning_rate": 1.863908546543277e-05,
      "loss": 1.1684,
      "step": 500
    },
    {
      "epoch": 0.54,
      "learning_rate": 1.7278170930865544e-05,
      "loss": 0.6721,
      "step": 1000
    },
    {
      "epoch": 0.82,
      "learning_rate": 1.5917256396298314e-05,
      "loss": 0.6004,
      "step": 1500
    },
    {
      "epoch": 1.09,
      "learning_rate": 1.4556341861731085e-05,
      "loss": 0.5305,
      "step": 2000
    },
    {
      "epoch": 1.36,
      "learning_rate": 1.3195427327163855e-05,
      "loss": 0.4386,
      "step": 2500
    },
    {
      "epoch": 1.63,
      "learning_rate": 1.1834512792596626e-05,
      "loss": 0.4283,
      "step": 3000
    },
    {
      "epoch": 1.91,
      "learning_rate": 1.0473598258029396e-05,
      "loss": 0.4255,
      "step": 3500
    },
    {
      "epoch": 2.18,
      "learning_rate": 9.112683723462167e-06,
      "loss": 0.3535,
      "step": 4000
    },
    {
      "epoch": 2.45,
      "learning_rate": 7.751769188894938e-06,
      "loss": 0.308,
      "step": 4500
    },
    {
      "epoch": 2.72,
      "learning_rate": 6.390854654327709e-06,
      "loss": 0.3116,
      "step": 5000
    },
    {
      "epoch": 2.99,
      "learning_rate": 5.029940119760479e-06,
      "loss": 0.3116,
      "step": 5500
    },
    {
      "epoch": 3.27,
      "learning_rate": 3.66902558519325e-06,
      "loss": 0.2396,
      "step": 6000
    },
    {
      "epoch": 3.54,
      "learning_rate": 2.308111050626021e-06,
      "loss": 0.2356,
      "step": 6500
    },
    {
      "epoch": 3.81,
      "learning_rate": 9.471965160587916e-07,
      "loss": 0.2361,
      "step": 7000
    },
    {
      "epoch": 4.0,
      "step": 7348,
      "total_flos": 4.0425346910601216e+17,
      "train_loss": 0.4368036111282407,
      "train_runtime": 19719.8453,
      "train_samples_per_second": 26.824,
      "train_steps_per_second": 0.373
    }
  ],
  "max_steps": 7348,
  "num_train_epochs": 4,
  "total_flos": 4.0425346910601216e+17,
  "trial_name": null,
  "trial_params": null
}