checkpoint-520/trainer_state.json

{
  "best_metric": 0.6354775828460039,
  "best_model_checkpoint": "brand-safety-model\\checkpoint-325",
  "epoch": 8.0,
  "eval_steps": 500,
  "global_step": 520,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.7692307692307693,
      "grad_norm": 2.6331872940063477,
      "learning_rate": 2.7692307692307694e-05,
      "loss": 3.319,
      "step": 50
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.42300194931773877,
      "eval_loss": 2.7462074756622314,
      "eval_runtime": 1.5522,
      "eval_samples_per_second": 330.495,
      "eval_steps_per_second": 10.952,
      "step": 65
    },
    {
      "epoch": 1.5384615384615383,
      "grad_norm": 3.3225834369659424,
      "learning_rate": 2.5384615384615386e-05,
      "loss": 2.6272,
      "step": 100
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.5009746588693957,
      "eval_loss": 2.07950758934021,
      "eval_runtime": 1.523,
      "eval_samples_per_second": 336.824,
      "eval_steps_per_second": 11.162,
      "step": 130
    },
    {
      "epoch": 2.3076923076923075,
      "grad_norm": 3.2619457244873047,
      "learning_rate": 2.307692307692308e-05,
      "loss": 2.137,
      "step": 150
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.5769980506822612,
      "eval_loss": 1.6682791709899902,
      "eval_runtime": 1.5108,
      "eval_samples_per_second": 339.548,
      "eval_steps_per_second": 11.252,
      "step": 195
    },
    {
      "epoch": 3.076923076923077,
      "grad_norm": 3.448869228363037,
      "learning_rate": 2.076923076923077e-05,
      "loss": 1.7515,
      "step": 200
    },
    {
      "epoch": 3.8461538461538463,
      "grad_norm": 3.639822483062744,
      "learning_rate": 1.8461538461538465e-05,
      "loss": 1.469,
      "step": 250
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.6101364522417154,
      "eval_loss": 1.472076654434204,
      "eval_runtime": 1.5018,
      "eval_samples_per_second": 341.601,
      "eval_steps_per_second": 11.32,
      "step": 260
    },
    {
      "epoch": 4.615384615384615,
      "grad_norm": 4.488701820373535,
      "learning_rate": 1.6153846153846154e-05,
      "loss": 1.2405,
      "step": 300
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.6354775828460039,
      "eval_loss": 1.34968101978302,
      "eval_runtime": 1.5205,
      "eval_samples_per_second": 337.4,
      "eval_steps_per_second": 11.181,
      "step": 325
    },
    {
      "epoch": 5.384615384615385,
      "grad_norm": 3.624941825866699,
      "learning_rate": 1.3846153846153847e-05,
      "loss": 1.1023,
      "step": 350
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.6335282651072125,
      "eval_loss": 1.2936445474624634,
      "eval_runtime": 1.5136,
      "eval_samples_per_second": 338.922,
      "eval_steps_per_second": 11.231,
      "step": 390
    },
    {
      "epoch": 6.153846153846154,
      "grad_norm": 4.275500774383545,
      "learning_rate": 1.153846153846154e-05,
      "loss": 1.0241,
      "step": 400
    },
    {
      "epoch": 6.923076923076923,
      "grad_norm": 3.8007335662841797,
      "learning_rate": 9.230769230769232e-06,
      "loss": 0.9206,
      "step": 450
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.631578947368421,
      "eval_loss": 1.2855250835418701,
      "eval_runtime": 1.4953,
      "eval_samples_per_second": 343.071,
      "eval_steps_per_second": 11.369,
      "step": 455
    },
    {
      "epoch": 7.6923076923076925,
      "grad_norm": 4.490166187286377,
      "learning_rate": 6.923076923076923e-06,
      "loss": 0.8374,
      "step": 500
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.6354775828460039,
      "eval_loss": 1.2579463720321655,
      "eval_runtime": 1.5008,
      "eval_samples_per_second": 341.827,
      "eval_steps_per_second": 11.328,
      "step": 520
    }
  ],
  "logging_steps": 50,
  "max_steps": 650,
  "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": 648555989889024.0,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}