graphcodebert-base-lowLR-highBatchSize/checkpoint-350/config_hyperparams.json

{
  "train_config": {
    "model_name": "microsoft/graphcodebert-base",
    "num_epochs": 1,
    "batch_size": 256,
    "learning_rate": 1e-06,
    "max_length": 512,
    "num_labels": 2,
    "loss_type": "r-drop",
    "focal_alpha": 1.0,
    "focal_gamma": 2.0,
    "r_drop_alpha": 10.0,
    "infonce_temperature": 0.07,
    "infonce_weight": 0.5,
    "label_smoothing": 0.5,
    "adversarial_epsilon": 0.5,
    "use_swa": false,
    "swa_start_epoch": 0,
    "swa_lr": 1e-05,
    "data_augmentation": true,
    "aug_rename_prob": 0.8,
    "aug_format_prob": 0.8,
    "freeze_base": true,
    "seed": 42,
    "use_wandb": true,
    "mixup_alpha": 1.0,
    "low_pass_keep_ratio": 0.5,
    "freq_consistency_weight": 0.5
  },
  "training_arguments": {
    "output_dir": "output_checkpoints/graphcodebert-base-lowLR-highBatchSize/",
    "num_train_epochs": 1,
    "per_device_train_batch_size": 256,
    "per_device_eval_batch_size": 512,
    "learning_rate": 1e-06,
    "warmup_steps": 204,
    "weight_decay": 0.1,
    "logging_steps": 10,
    "eval_steps": 1000,
    "save_steps": 50,
    "metric_for_best_model": "macro_f1",
    "greater_is_better": true,
    "save_total_limit": 4,
    "fp16": true,
    "seed": 42
  },
  "training_state": {
    "global_step": 350,
    "epoch": 0.3424657534246575,
    "best_metric": null,
    "best_model_checkpoint": null
  }
}