trainer_state.json

{
  "best_metric": 0.2894308269023895,
  "best_model_checkpoint": "xlm-roberta-base-ner/checkpoint-1257",
  "epoch": 5.0,
  "eval_steps": 500,
  "global_step": 6285,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.39777247414478917,
      "grad_norm": 10.322513580322266,
      "learning_rate": 1.9204455051710422e-05,
      "loss": 0.5382,
      "step": 500
    },
    {
      "epoch": 0.7955449482895783,
      "grad_norm": 6.790235996246338,
      "learning_rate": 1.8408910103420846e-05,
      "loss": 0.395,
      "step": 1000
    },
    {
      "epoch": 1.0,
      "eval_f1": 0.8085669291338583,
      "eval_loss": 0.2894308269023895,
      "eval_precision": 0.8323174685514201,
      "eval_recall": 0.7861342479176874,
      "eval_runtime": 14.4214,
      "eval_samples_per_second": 150.817,
      "eval_steps_per_second": 18.861,
      "step": 1257
    },
    {
      "epoch": 1.1933174224343674,
      "grad_norm": 2.2459216117858887,
      "learning_rate": 1.7613365155131266e-05,
      "loss": 0.3372,
      "step": 1500
    },
    {
      "epoch": 1.5910898965791567,
      "grad_norm": 3.4064340591430664,
      "learning_rate": 1.681782020684169e-05,
      "loss": 0.3252,
      "step": 2000
    },
    {
      "epoch": 1.988862370723946,
      "grad_norm": 2.284067392349243,
      "learning_rate": 1.602227525855211e-05,
      "loss": 0.3229,
      "step": 2500
    },
    {
      "epoch": 2.0,
      "eval_f1": 0.8084385006592578,
      "eval_loss": 0.30411699414253235,
      "eval_precision": 0.8293185624114389,
      "eval_recall": 0.7885840274375306,
      "eval_runtime": 9.1696,
      "eval_samples_per_second": 237.197,
      "eval_steps_per_second": 29.663,
      "step": 2514
    },
    {
      "epoch": 2.386634844868735,
      "grad_norm": 3.480130672454834,
      "learning_rate": 1.5226730310262532e-05,
      "loss": 0.2692,
      "step": 3000
    },
    {
      "epoch": 2.7844073190135243,
      "grad_norm": 9.41341781616211,
      "learning_rate": 1.4431185361972953e-05,
      "loss": 0.2692,
      "step": 3500
    },
    {
      "epoch": 3.0,
      "eval_f1": 0.7990613807583056,
      "eval_loss": 0.3256344497203827,
      "eval_precision": 0.8057285180572852,
      "eval_recall": 0.7925036746692797,
      "eval_runtime": 9.1536,
      "eval_samples_per_second": 237.61,
      "eval_steps_per_second": 29.715,
      "step": 3771
    },
    {
      "epoch": 3.1821797931583133,
      "grad_norm": 10.631413459777832,
      "learning_rate": 1.3635640413683375e-05,
      "loss": 0.2578,
      "step": 4000
    },
    {
      "epoch": 3.579952267303103,
      "grad_norm": 1.1227880716323853,
      "learning_rate": 1.2840095465393797e-05,
      "loss": 0.2237,
      "step": 4500
    },
    {
      "epoch": 3.977724741447892,
      "grad_norm": 3.822526693344116,
      "learning_rate": 1.2044550517104217e-05,
      "loss": 0.2554,
      "step": 5000
    },
    {
      "epoch": 4.0,
      "eval_f1": 0.8008445631248836,
      "eval_loss": 0.34640657901763916,
      "eval_precision": 0.8121929714069782,
      "eval_recall": 0.7898089171974523,
      "eval_runtime": 9.0831,
      "eval_samples_per_second": 239.454,
      "eval_steps_per_second": 29.946,
      "step": 5028
    },
    {
      "epoch": 4.375497215592681,
      "grad_norm": 15.420336723327637,
      "learning_rate": 1.1249005568814639e-05,
      "loss": 0.2123,
      "step": 5500
    },
    {
      "epoch": 4.77326968973747,
      "grad_norm": 2.319063663482666,
      "learning_rate": 1.045346062052506e-05,
      "loss": 0.2157,
      "step": 6000
    },
    {
      "epoch": 5.0,
      "eval_f1": 0.8262004492199356,
      "eval_loss": 0.312337189912796,
      "eval_precision": 0.8189914550487424,
      "eval_recall": 0.8335374816266536,
      "eval_runtime": 9.0172,
      "eval_samples_per_second": 241.205,
      "eval_steps_per_second": 30.164,
      "step": 6285
    }
  ],
  "logging_steps": 500,
  "max_steps": 12570,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 10,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 1621809589393752.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}