Shaun ZhouandClaude Opus 5
86c86b105c
实际工作流是：simclr 训练 -> 聚类 -> 人工归类 -> 分类训练 -> GradCAM/SHAP 生成
热力图与 box。据此移除不参与其中的检测/定位链路，共 31 个文件、约 2600 行：
- 模型：defect_net、postprocess、necks/、heads/center_head
- 数据：film_defect（YOLO 读取端）、target_encoder
- 引擎：trainer、evaluator、losses
- 入口：train.py、eval.py
- 脚本：convert_annotations、split_dataset、export_onnx、profile_inference
- 配置：configs/{train,eval,model,dataset}
- utils/labels（多类/单类标签模式，仅检测分类头使用）
共享模块按实际用量瘦身：engine/metrics 只留 AverageMeter（检测指标随之移除）；
datasets/yolo 只留写出侧（IMAGE_SUFFIXES 与 write_data_yaml）——第 5 步产出的
YOLO 标签是给外部检测训练用的交换格式，读取端不再需要。
裁剪过程中修掉一个真 bug：generate_boxes 的 --stage 默认 6（layer3）对 tiny_conv
这类非 nn.Sequential 骨架直接抛异常。现在自动退回骨架最终输出并警告，
另加 --stage -1 显式选择最终输出。
五步工作流已端到端实跑验证（simclr 1 轮 -> 聚类 -> 导出 -> 分类 1 轮 ->
gradcam 与 shap 两种方法各生成一遍）。
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
72 lines
2.7 KiB
Python
72 lines
2.7 KiB
Python
"""卷积骨架与分类模型。"""
|
|
|
|
import pytest
|
|
|
|
torch = pytest.importorskip("torch")
|
|
|
|
from models import build_model # noqa: E402
|
|
from models.backbones.resnet import _SUPPORTED as _SUPPORTED_ARCHS # noqa: E402
|
|
from models.heads.projection import ProjectionHead # noqa: E402
|
|
|
|
|
|
@pytest.mark.parametrize("arch", sorted(_SUPPORTED_ARCHS))
|
|
def test_every_supported_torchvision_arch_actually_works(arch):
|
|
"""清单里列出的每个骨架都必须真能前向，且声明的 out_channels 与实际一致。
|
|
|
|
回归测试：shufflenet_v2_x1_0 曾被列为支持，但它没有 `.features` 属性，
|
|
构造时直接 AttributeError。
|
|
"""
|
|
from models.registry import BACKBONES
|
|
|
|
backbone = BACKBONES.build({"name": "torchvision", "arch": arch, "pretrained": False}).eval()
|
|
with torch.no_grad():
|
|
feat = backbone(torch.randn(1, 3, 64, 64))
|
|
|
|
assert feat.dim() == 4
|
|
assert feat.shape[1] == backbone.out_channels == _SUPPORTED_ARCHS[arch]
|
|
assert backbone.out_stride == 32
|
|
|
|
|
|
def test_unsupported_arch_is_rejected():
|
|
from models.registry import BACKBONES
|
|
|
|
with pytest.raises(ValueError, match="不支持的骨架"):
|
|
BACKBONES.build({"name": "torchvision", "arch": "vit_b_16"})
|
|
|
|
|
|
def test_tiny_conv_shapes():
|
|
from models.registry import BACKBONES
|
|
|
|
backbone = BACKBONES.build(
|
|
{"name": "tiny_conv", "widths": [8, 16, 32], "depths": [1, 1, 1]}
|
|
).eval()
|
|
with torch.no_grad():
|
|
feat = backbone(torch.randn(2, 3, 64, 64))
|
|
assert feat.shape[:2] == (2, 32)
|
|
assert backbone.out_channels == 32
|
|
# stem 下采样 2 倍，其后每个 stage(除第一个)各下采样 2 倍
|
|
assert backbone.out_stride == 8
|
|
|
|
|
|
@pytest.mark.parametrize("arch", ["resnet18", "shufflenet_v2_x1_0", "mobilenet_v3_small"])
|
|
def test_alternative_backbones_drive_simclr(arch):
|
|
"""换骨架后 SimCLR 的投影头要能接上（out_channels 必须准）。"""
|
|
from models.registry import BACKBONES
|
|
|
|
backbone = BACKBONES.build({"name": "torchvision", "arch": arch, "pretrained": False}).eval()
|
|
head = ProjectionHead(backbone.out_channels, 32, 16)
|
|
with torch.no_grad():
|
|
assert head(backbone(torch.randn(2, 3, 64, 64))).shape == (2, 16)
|
|
|
|
|
|
def test_classifier_shapes_and_convolution_only():
|
|
model = build_model({
|
|
"name": "classifier", "num_classes": 4,
|
|
"backbone": {"name": "tiny_conv", "widths": [8, 16], "depths": [1, 1]},
|
|
"classification": {"dropout": 0.1},
|
|
}).eval()
|
|
with torch.no_grad():
|
|
assert model(torch.randn(3, 3, 64, 64)).shape == (3, 4)
|
|
forbidden = (torch.nn.MultiheadAttention, torch.nn.Transformer, torch.nn.LSTM)
|
|
assert not any(isinstance(m, forbidden) for m in model.modules())
|