feat: 增加全局config

brep2sdf/config/default_config.py

@@ -0,0 +1,102 @@
+from dataclasses import dataclass
+from typing import Tuple, Optional
+
+@dataclass
+class ModelConfig:
+    """模型相关配置"""
+    brep_feature_dim: int = 48
+    use_cf: bool = True
+    embed_dim: int = 768
+    latent_dim: int = 256
+    
+    # 点云采样配置
+    num_surf_points: int = 16  # 每个面采样点数
+    num_edge_points: int = 4   # 每条边采样点数
+
+@dataclass
+class DataConfig:
+    """数据相关配置"""
+    max_face: int = 70
+    max_edge: int = 70
+    bbox_scaled: float = 1.0
+    
+    # 数据路径
+    brep_dir: str = '/home/wch/brep2sdf/test_data/pkl'
+    sdf_dir: str = '/home/wch/brep2sdf/test_data/sdf'
+    valid_data_dir: str = '/home/wch/brep2sdf/test_data/result/pkl'
+    save_dir: str = 'checkpoints'
+
+@dataclass
+class TrainConfig:
+    """训练相关配置"""
+    # 基本训练参数
+    batch_size: int = 32
+    num_workers: int = 4
+    num_epochs: int = 100
+    learning_rate: float = 1e-4
+    min_lr: float = 1e-6
+    weight_decay: float = 0.01
+    
+    # 梯度和损失相关
+    grad_weight: float = 0.1
+    max_grad_norm: float = 1.0
+    
+    # 学习率调度器参数
+    lr_scheduler: str = 'cosine'  # ['cosine', 'linear', 'step']
+    warmup_epochs: int = 5
+    
+    # 保存和验证
+    save_freq: int = 10  # 每多少个epoch保存一次
+    val_freq: int = 1    # 每多少个epoch验证一次
+
+@dataclass
+class LogConfig:
+    """日志相关配置"""
+    # wandb配置
+    use_wandb: bool = True
+    project_name: str = 'brep2sdf'
+    run_name: str = 'training_run'
+    log_interval: int = 10
+    
+    # 本地日志
+    log_dir: str = 'logs'
+    log_level: str = 'INFO'
+
+@dataclass
+class Config:
+    """总配置类"""
+    model: ModelConfig = ModelConfig()
+    data: DataConfig = DataConfig()
+    train: TrainConfig = TrainConfig()
+    log: LogConfig = LogConfig()
+    
+    def __post_init__(self):
+        """初始化后的处理"""
+        # 可以在这里添加配置验证逻辑
+        pass
+
+    @classmethod
+    def from_dict(cls, config_dict: dict) -> 'Config':
+        """从字典创建配置"""
+        model_config = ModelConfig(**config_dict.get('model', {}))
+        data_config = DataConfig(**config_dict.get('data', {}))
+        train_config = TrainConfig(**config_dict.get('train', {}))
+        log_config = LogConfig(**config_dict.get('log', {}))
+        
+        return cls(
+            model=model_config,
+            data=data_config,
+            train=train_config,
+            log=log_config
+        )
+
+def get_default_config() -> Config:
+    """获取默认配置"""
+    return Config()
+
+def load_config(config_path: str) -> Config:
+    """从文件加载配置"""
+    import yaml
+    with open(config_path, 'r') as f:
+        config_dict = yaml.safe_load(f)
+    return Config.from_dict(config_dict)