增加注释，去掉dropout

classifier/embedding_food_classifier_app.py

@@ -101,12 +101,12 @@ class EmbeddingFoodClassifierApp:
                 self.faiss_index = faiss.read_index(index_path)
                 print(f"FAISS索引已加载: {self.faiss_index.ntotal} 个向量")
                 
-                # 加载图片路径映射
+                # 加载图片路径映射，就是一个list,一张张图片的路径
                 paths_path = os.path.join(index_dir, 'image_paths.pkl')
                 with open(paths_path, 'rb') as f:
                     self.image_paths = pickle.load(f)
                 
-                # 加载标签映射
+                # 加载标签映射，就是一个list,一张张图片的分类，不过是0，1，2，3，4这种
                 labels_path = os.path.join(index_dir, 'labels.pkl')
                 with open(labels_path, 'rb') as f:
                     self.labels = pickle.load(f)
@@ -605,8 +605,11 @@ class EmbeddingFoodClassifierApp:
                 if i < len(indices[0]) and indices[0][i] < len(self.labels):
                     idx = indices[0][i]
                     score = scores[0][i]
+                    # 根据FAISS索引拿到对应的分类id,0,1,2,3,4
                     class_idx = self.labels[idx]
+                    # 根据分类id拿到对应的中文分类名称
                     class_name = self.class_names[class_idx]
+                    # 根据FAISS的索引，拿到图片原始的路径
                     image_path = self.image_paths[idx]
                     
                     similar_classes.append(class_name)

net/resnet_embedding.py

@@ -42,11 +42,17 @@ class ResNet50EmbeddingNet(nn.Module):
         self.backbone.fc = nn.Identity()
         
         # 添加embedding层
+        # self.embedding_layer = nn.Sequential(
+        #     nn.Linear(backbone_output_dim, embedding_dim),
+        #     nn.BatchNorm1d(embedding_dim),
+        #     nn.ReLU(inplace=True),
+        #     nn.Dropout(0.2)
+        # )
+        # 暂时不用加Dropout
         self.embedding_layer = nn.Sequential(
             nn.Linear(backbone_output_dim, embedding_dim),
             nn.BatchNorm1d(embedding_dim),
-            nn.ReLU(inplace=True),
+            nn.ReLU(inplace=True)
-            nn.Dropout(0.2)
         )
         
         # 图片预处理变换（用于推理）