增加一些设置。

.gitignore

@@ -2,7 +2,8 @@
 /.idea/
 /model/
 /faiss_vector_db/demo_faiss_index/
-/faiss_vector_db/faiss_index/
+/faiss_vector_db/DishClassification/faiss_index/
-/faiss_vector_db/faiss_index092901/
+/faiss_vector_db/DishClassification/faiss_index092901/
-/faiss_vector_db/faiss_index101001/
+/faiss_vector_db/DishClassification/faiss_index101001/
 /faiss_vector_db/faiss_index*/
+/faiss_vector_db/WholeIngredientRecognition/faiss_index*/

build_faiss_index.py

@@ -470,7 +470,7 @@ def main():
     # 配置参数
     MODEL_PATH = "model/embedding_20250930_102826/best_embedding_model.pth"
     TRAIN_DIR = "dataset/train"
-    OUTPUT_DIR = "faiss_vector_db/faiss_index"
+    OUTPUT_DIR = "faiss_vector_db/DishClassification/faiss_index"
     BATCH_SIZE = 16
     INDEX_TYPE = 'flat'  # 'flat', 'ivf', 'hnsw'
     EMBEDDING_DIM = 512

classifier/embedding_food_classifier_app.py

@@ -64,11 +64,15 @@ class EmbeddingFoodClassifierApp:
             # 获取当前脚本所在目录
             BASE_DIR = os.path.dirname(os.path.abspath(__file__))
             # model_path = "../model/embedding_20251011_133653/best_embedding_model.pth"
-            model_path = os.path.join(BASE_DIR, "../model/embedding_20251011_133653/best_embedding_model.pth")
+            # model_path = os.path.join(BASE_DIR, "../model/embedding_20251011_133653/best_embedding_model.pth")
+            model_path = os.path.join(BASE_DIR, "../model/WholeIngredientRecognition/embedding_20251017_145836/best_embedding_model.pth")
+            # model_path = os.path.join(BASE_DIR, "../model/DishClassification/embedding_20251011_133653/best_embedding_model.pth")
 
             # FAISS索引目录
             # index_dir = "../faiss_vector_db/faiss_index"
-            index_dir = os.path.join(BASE_DIR, "../faiss_vector_db/faiss_index")
+            # index_dir = os.path.join(BASE_DIR, "../faiss_vector_db/faiss_index")
+            index_dir = os.path.join(BASE_DIR, "../faiss_vector_db/WholeIngredientRecognition/faiss_index")
+            # index_dir = os.path.join(BASE_DIR, "../faiss_vector_db/DishClassification/faiss_index")
 
             if os.path.exists(model_path) and os.path.exists(index_dir):
                 # 1. 加载embedding模型

classifier/embedding_food_classifier_search_app.py

@@ -759,7 +759,7 @@ class EmbeddingFoodClassifierApp:
     def save_updated_index(self):
         """保存更新后的索引和元数据"""
         try:
-            index_dir = "../faiss_vector_db/faiss_index"
+            index_dir = "../faiss_vector_db/DishClassification/faiss_index"
             
             # 保存向量库为 embeddings.json（list[list[float]]）
             emb_path = os.path.join(index_dir, 'embeddings.json')

faiss_vector_db/README_FAISS.md

@@ -111,7 +111,7 @@ builder = FAISSIndexBuilder(
 # 构建完整索引
 index = builder.build_complete_index(
     train_dir="../dataset/train",
-    output_dir="faiss_index",
+    output_dir="DishClassification/faiss_index",
     batch_size=16,
     index_type='flat'  # 'flat', 'ivf', 'hnsw'
 )

faiss_vector_db/build_faiss_index.py

@@ -133,6 +133,8 @@ class FAISSIndexBuilder:
             print(f"类别 '{class_name}': {len(image_files)} 张图片")
             
             for image_file in image_files:
+                # 可以不添加那些增强的图片
+                if image_file.startswith("img"):
                     image_path = os.path.join(class_dir, image_file)
                     image_paths.append(image_path)
                     class_names.append(class_name)
@@ -477,9 +479,13 @@ class FAISSSearcher:
 def main():
     """主函数"""
     # 配置参数
-    MODEL_PATH = "../model/embedding_20251011_133653/best_embedding_model.pth"
+    # MODEL_PATH = "../model/embedding_20251011_133653/best_embedding_model.pth"
-    TRAIN_DIR = "../dataset/train"
+    MODEL_PATH = "../model/WholeIngredientRecognition/embedding_20251017_145836/best_embedding_model.pth"
-    OUTPUT_DIR = "faiss_index"
+    # MODEL_PATH = "../model/DishClassification/embedding_20251011_133653/best_embedding_model.pth"
+    TRAIN_DIR = "../dataset/WholeIngredientRecognition/train"
+    # TRAIN_DIR = "../dataset/DishClassification/train"
+    OUTPUT_DIR = "WholeIngredientRecognition/faiss_index"
+    # OUTPUT_DIR = "DishClassification/faiss_index"
     BATCH_SIZE = 16
     INDEX_TYPE = 'flat'  # 'flat', 'ivf', 'hnsw'
     EMBEDDING_DIM = 512

faiss_vector_db/run_faiss_demo.py

@@ -87,7 +87,7 @@ def run_search_demo():
         from faiss_vector_db.build_faiss_index import FAISSSearcher
         
         # 配置参数
-        index_dir = "faiss_index"
+        index_dir = "DishClassification/faiss_index"
         model_path = "../model/embedding_20250917_145342/best_embedding_model.pth"
         
         # 检查索引是否存在
@@ -206,7 +206,7 @@ def main():
     print("✓ 文件检查通过")
     
     # 检查是否已有索引
-    index_exists = os.path.exists("faiss_index/faiss_index.bin")
+    index_exists = os.path.exists("DishClassification/faiss_index/faiss_index.bin")
     
     if index_exists:
         print("✓ 发现已存在的FAISS索引")

faiss_vector_db/visualize_embeddings.py

@@ -162,8 +162,10 @@ def plot_2d(Z: np.ndarray, y: np.ndarray, title: str, out_path: Optional[str] =
 
 def main():
     parser = argparse.ArgumentParser(description="可视化高维 embedding 并进行坍塌诊断")
-    parser.add_argument("--embeddings", type=str, default=os.path.join( "faiss_index", "embeddings.json"), help="embeddings.json 路径")
+    # parser.add_argument("--embeddings", type=str, default=os.path.join("DishClassification/faiss_index", "embeddings.json"), help="embeddings.json 路径")
-    parser.add_argument("--labels", type=str, default=os.path.join("faiss_index", "labels.json"), help="labels.json 路径")
+    parser.add_argument("--embeddings", type=str, default=os.path.join("WholeIngredientRecognition/faiss_index", "embeddings.json"), help="embeddings.json 路径")
+    # parser.add_argument("--labels", type=str, default=os.path.join("DishClassification/faiss_index", "labels.json"), help="labels.json 路径")
+    parser.add_argument("--labels", type=str, default=os.path.join("WholeIngredientRecognition/faiss_index", "labels.json"), help="labels.json 路径")
     parser.add_argument("--method", type=str, default="pca", choices=["pca", "tsne", "umap"], help="降维方法")
     parser.add_argument("--seed", type=int, default=42)
     parser.add_argument("--max_points", type=int, default=None, help="抽样上限，避免t-SNE/UMAP过慢；None为全量")

toAndroid/toAndroidSearch.py

@@ -197,8 +197,8 @@ def build_and_export_similarity_head(
 
 def main():
     print("开始转换检索头为移动端格式（TorchScript）...")
-    embeddings_path = "../faiss_vector_db/faiss_index/embeddings.json"
+    embeddings_path = "../faiss_vector_db/DishClassification/faiss_index/embeddings.json"
-    output_path = "../faiss_vector_db/faiss_index/similarity_head.pt"
+    output_path = "../faiss_vector_db/DishClassification/faiss_index/similarity_head.pt"
 
     try:
         path, n_items = build_and_export_similarity_head(embeddings_path, output_path)

train/train_embedding.py

@@ -660,7 +660,7 @@ def main(task_key: str = "dish"):
         batch_size=BATCH_SIZE, 
         shuffle=False, 
         num_workers=0,
-        drop_last=True
+        drop_last=False
     )
     
     logger.info(f"训练集大小: {len(train_dataset)}")
@@ -811,4 +811,4 @@ if __name__ == "__main__":
     parser = argparse.ArgumentParser()
     parser.add_argument("task", choices=list(TASKS.keys()), nargs="?", default="dish")
     args = parser.parse_args()
-    main(args.task)
+    main("whole_ingredient")