$ loading_
提供 PyTorch 深度学习模式与最佳实践，帮助构建高效可复现的训练流程。
复制安装指令，让 AI 自动完成配置 · 推荐新手
请帮我安装 askskill 上的 "pytorch-patterns" 技能： 1. 下载 https://raw.githubusercontent.com/affaan-m/ECC/main/skills/pytorch-patterns/SKILL.md 2. 保存为 ~/.claude/skills/pytorch-patterns/SKILL.md 3. 装好后重载技能，告诉我可以用了
请给我一套基于 PyTorch 的训练管线最佳实践，包含项目目录结构、配置管理、随机种子设置、训练与验证循环、checkpoint 保存、日志记录和可复现性建议。
一份结构化的训练管线方案，涵盖核心模块设计与可复现实现建议。
我正在用 PyTorch 训练图像模型，请总结 DataLoader、Dataset、数据增强、多进程加载、pin_memory 和预取策略的最佳实践，并指出常见性能瓶颈与排查方法。
一套数据加载优化建议，包含性能调优点、常见错误和排查清单。
请基于 PyTorch 最佳实践，说明如何设计可维护的模型架构代码，包括模块拆分、forward 规范、初始化策略、设备管理、混合精度训练和实验追踪建议。
一份模型工程规范建议，帮助提升代码可维护性、训练稳定性与实验管理能力。
Idiomatic PyTorch patterns and best practices for building robust, efficient, and reproducible deep learning applications.
Always write code that works on both CPU and GPU without hardcoding devices.
# Good: Device-agnostic
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
model = MyModel().to(device)
data = data.to(device)
# Bad: Hardcoded device
model = MyModel().cuda() # Crashes if no GPU
data = data.cuda()
Set all random seeds for reproducible results.
# Good: Full reproducibility setup
def set_seed(seed: int = 42) -> None:
torch.manual_seed(seed)
torch.cuda.manual_seed_all(seed)
np.random.seed(seed)
random.seed(seed)
torch.backends.cudnn.deterministic = True
torch.backends.cudnn.benchmark = False
# Bad: No seed control
model = MyModel() # Different weights every run
Always document and verify tensor shapes.
# Good: Shape-annotated forward pass
def forward(self, x: torch.Tensor) -> torch.Tensor:
# x: (batch_size, channels, height, width)
x = self.conv1(x) # -> (batch_size, 32, H, W)
x = self.pool(x) # -> (batch_size, 32, H//2, W//2)
x = x.view(x.size(0), -1) # -> (batch_size, 32*H//2*W//2)
return self.fc(x) # -> (batch_size, num_classes)
# Bad: No shape tracking
def forward(self, x):
x = self.conv1(x)
x = self.pool(x)
x = x.view(x.size(0), -1) # What size is this?
return self.fc(x) # Will this even work?
# Good: Well-organized module
class ImageClassifier(nn.Module):
def __init__(self, num_classes: int, dropout: float = 0.5) -> None:
super().__init__()
self.features = nn.Sequential(
nn.Conv2d(3, 64, kernel_size=3, padding=1),
nn.BatchNorm2d(64),
nn.ReLU(inplace=True),
nn.MaxPool2d(2),
)
self.classifier = nn.Sequential(
nn.Dropout(dropout),
nn.Linear(64 * 16 * 16, num_classes),
)
def forward(self, x: torch.Tensor) -> torch.Tensor:
x = self.features(x)
x = x.view(x.size(0), -1)
return self.classifier(x)
# Bad: Everything in forward
class ImageClassifier(nn.Module):
def __init__(self):
super().__init__()
def forward(self, x):
x = F.conv2d(x, weight=self.make_weight()) # Creates weight each call!
return x
# Good: Explicit initialization
def _init_weights(self, module: nn.Module) -> None:
if isinstance(module, nn.Linear):
nn.init.kaiming_normal_(module.weight, mode="fan_out", nonlinearity="relu")
if module.bias is not None:
nn.init.zeros_(module.bias)
elif isinstance(module, nn.Conv2d):
nn.init.kaiming_normal_(module.weight, mode="fan_out", nonlinearity="relu")
elif isinstance(module, nn.BatchNorm2d):
nn.init.ones_(module.weight)
nn.init.zeros_(module.bias)
model = MyModel()
model.apply(model._init_weights)
# Good: Complete training loop with best practices
def train_one_epoch(
model: nn.Module,
dataloader: DataLoader,
optimizer: torch.optim.Optimizer,
criterion: nn.Module,
device: torch.device,
scaler: torch.amp.GradScaler | None = None,
) -> float:
model.train() # Always set train mode
total_loss = 0.0
for batch_idx, (data, target) in enumerate(dataloader):
data, target = data.to(device), target.to(device)
…
通过双评审智能体对结果进行对抗式校验，提升输出发布前的可靠性
帮助开发者掌握 Django 架构模式、DRF 设计与生产级实践