checkpoint-6/trainer_state.json

{
  "best_metric": 10.3258695602417,
  "best_model_checkpoint": "finetuned_models/qwen-32b-multi-stock-test_202603191229/checkpoint-6",
  "epoch": 2.1333333333333333,
  "eval_steps": 2,
  "global_step": 6,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.35555555555555557,
      "grad_norm": 4.793458461761475,
      "learning_rate": 5e-05,
      "loss": 16.1288,
      "step": 1
    },
    {
      "epoch": 0.7111111111111111,
      "grad_norm": 5.042018413543701,
      "learning_rate": 4e-05,
      "loss": 16.6333,
      "step": 2
    },
    {
      "epoch": 0.7111111111111111,
      "eval_loss": 15.579117774963379,
      "eval_runtime": 59.8755,
      "eval_samples_per_second": 0.401,
      "eval_steps_per_second": 0.401,
      "step": 2
    },
    {
      "epoch": 1.0666666666666667,
      "grad_norm": 5.352873802185059,
      "learning_rate": 3e-05,
      "loss": 16.0459,
      "step": 3
    },
    {
      "epoch": 1.4222222222222223,
      "grad_norm": 6.826162815093994,
      "learning_rate": 2e-05,
      "loss": 14.5282,
      "step": 4
    },
    {
      "epoch": 1.4222222222222223,
      "eval_loss": 13.443461418151855,
      "eval_runtime": 59.3883,
      "eval_samples_per_second": 0.404,
      "eval_steps_per_second": 0.404,
      "step": 4
    },
    {
      "epoch": 1.7777777777777777,
      "grad_norm": 16.21648406982422,
      "learning_rate": 1e-05,
      "loss": 13.3791,
      "step": 5
    },
    {
      "epoch": 2.1333333333333333,
      "grad_norm": 32.58063507080078,
      "learning_rate": 0.0,
      "loss": 11.6245,
      "step": 6
    },
    {
      "epoch": 2.1333333333333333,
      "eval_loss": 10.3258695602417,
      "eval_runtime": 58.4873,
      "eval_samples_per_second": 0.41,
      "eval_steps_per_second": 0.41,
      "step": 6
    }
  ],
  "logging_steps": 1,
  "max_steps": 6,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 2,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 6.709791546189619e+16,
  "train_batch_size": 1,
  "trial_name": null,
  "trial_params": null
}