Spaces:
Runtime error
Runtime error
Update app.py
#2
by
JoJosmin
- opened
app.py
CHANGED
@@ -9,6 +9,7 @@ import json
|
|
9 |
import numpy as np
|
10 |
from ultralytics import YOLO
|
11 |
import cv2
|
|
|
12 |
|
13 |
# Load CLIP model and tokenizer
|
14 |
@st.cache_resource
|
@@ -28,14 +29,6 @@ def load_yolo_model():
|
|
28 |
|
29 |
yolo_model = load_yolo_model()
|
30 |
|
31 |
-
# Load and process data
|
32 |
-
@st.cache_data
|
33 |
-
def load_data():
|
34 |
-
with open('./musinsa-final.json', 'r', encoding='utf-8') as f:
|
35 |
-
return json.load(f)
|
36 |
-
|
37 |
-
data = load_data()
|
38 |
-
|
39 |
# Helper functions
|
40 |
def load_image_from_url(url, max_retries=3):
|
41 |
for attempt in range(max_retries):
|
@@ -49,6 +42,9 @@ def load_image_from_url(url, max_retries=3):
|
|
49 |
time.sleep(1)
|
50 |
else:
|
51 |
return None
|
|
|
|
|
|
|
52 |
|
53 |
def get_image_embedding(image):
|
54 |
image_tensor = preprocess_val(image).unsqueeze(0).to(device)
|
@@ -57,37 +53,6 @@ def get_image_embedding(image):
|
|
57 |
image_features /= image_features.norm(dim=-1, keepdim=True)
|
58 |
return image_features.cpu().numpy()
|
59 |
|
60 |
-
@st.cache_data
|
61 |
-
def process_database():
|
62 |
-
database_embeddings = []
|
63 |
-
database_info = []
|
64 |
-
|
65 |
-
for item in data:
|
66 |
-
image_url = item['이미지 링크'][0]
|
67 |
-
image = load_image_from_url(image_url)
|
68 |
-
if image is not None:
|
69 |
-
embedding = get_image_embedding(image)
|
70 |
-
database_embeddings.append(embedding)
|
71 |
-
database_info.append({
|
72 |
-
'id': item['\ufeff상품 ID'],
|
73 |
-
'category': item['카테고리'],
|
74 |
-
'brand': item['브랜드명'],
|
75 |
-
'name': item['제품명'],
|
76 |
-
'price': item['정가'],
|
77 |
-
'discount': item['할인율'],
|
78 |
-
'image_url': image_url
|
79 |
-
})
|
80 |
-
else:
|
81 |
-
st.warning(f"Skipping item {item['상품 ID']} due to image loading failure")
|
82 |
-
|
83 |
-
if database_embeddings:
|
84 |
-
return np.vstack(database_embeddings), database_info
|
85 |
-
else:
|
86 |
-
st.error("No valid embeddings were generated.")
|
87 |
-
return None, None
|
88 |
-
|
89 |
-
database_embeddings, database_info = process_database()
|
90 |
-
|
91 |
def get_text_embedding(text):
|
92 |
text_tokens = tokenizer([text]).to(device)
|
93 |
with torch.no_grad():
|
@@ -95,17 +60,33 @@ def get_text_embedding(text):
|
|
95 |
text_features /= text_features.norm(dim=-1, keepdim=True)
|
96 |
return text_features.cpu().numpy()
|
97 |
|
98 |
-
def
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
99 |
similarities = np.dot(database_embeddings, query_embedding.T).squeeze()
|
100 |
top_indices = np.argsort(similarities)[::-1][:top_k]
|
|
|
|
|
|
|
|
|
|
|
101 |
results = []
|
102 |
-
for idx in
|
103 |
results.append({
|
104 |
-
'info':
|
105 |
-
'similarity': similarities[idx]
|
106 |
})
|
107 |
return results
|
108 |
|
|
|
|
|
109 |
def detect_clothing(image):
|
110 |
results = yolo_model(image)
|
111 |
detections = results[0].boxes.data.cpu().numpy()
|
@@ -182,7 +163,7 @@ elif st.session_state.step == 'show_results':
|
|
182 |
cropped_image = crop_image(st.session_state.query_image, selected_detection['bbox'])
|
183 |
st.image(cropped_image, caption="Cropped Image", use_column_width=True)
|
184 |
query_embedding = get_image_embedding(cropped_image)
|
185 |
-
similar_images = find_similar_images(query_embedding)
|
186 |
|
187 |
st.subheader("Similar Items:")
|
188 |
for img in similar_images:
|
@@ -208,7 +189,7 @@ else: # Text search
|
|
208 |
if st.button("Search by Text"):
|
209 |
if query_text:
|
210 |
text_embedding = get_text_embedding(query_text)
|
211 |
-
similar_images = find_similar_images(text_embedding)
|
212 |
st.subheader("Similar Items:")
|
213 |
for img in similar_images:
|
214 |
col1, col2 = st.columns(2)
|
|
|
9 |
import numpy as np
|
10 |
from ultralytics import YOLO
|
11 |
import cv2
|
12 |
+
import chromadb
|
13 |
|
14 |
# Load CLIP model and tokenizer
|
15 |
@st.cache_resource
|
|
|
29 |
|
30 |
yolo_model = load_yolo_model()
|
31 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
32 |
# Helper functions
|
33 |
def load_image_from_url(url, max_retries=3):
|
34 |
for attempt in range(max_retries):
|
|
|
42 |
time.sleep(1)
|
43 |
else:
|
44 |
return None
|
45 |
+
#Load chromaDB
|
46 |
+
client = chromadb.PersistentClient(path="./clothesDB")
|
47 |
+
collection = client.get_collection(name="fashion_items_ver2")
|
48 |
|
49 |
def get_image_embedding(image):
|
50 |
image_tensor = preprocess_val(image).unsqueeze(0).to(device)
|
|
|
53 |
image_features /= image_features.norm(dim=-1, keepdim=True)
|
54 |
return image_features.cpu().numpy()
|
55 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
56 |
def get_text_embedding(text):
|
57 |
text_tokens = tokenizer([text]).to(device)
|
58 |
with torch.no_grad():
|
|
|
60 |
text_features /= text_features.norm(dim=-1, keepdim=True)
|
61 |
return text_features.cpu().numpy()
|
62 |
|
63 |
+
def get_all_embeddings_from_collection(collection):
|
64 |
+
all_embeddings = collection.get(include=['embeddings'])['embeddings']
|
65 |
+
return np.array(all_embeddings)
|
66 |
+
|
67 |
+
def get_metadata_from_ids(collection, ids):
|
68 |
+
results = collection.get(ids=ids)
|
69 |
+
return results['metadatas']
|
70 |
+
|
71 |
+
def find_similar_images(query_embedding, collection, top_k=5):
|
72 |
+
database_embeddings = get_all_embeddings_from_collection(collection)
|
73 |
similarities = np.dot(database_embeddings, query_embedding.T).squeeze()
|
74 |
top_indices = np.argsort(similarities)[::-1][:top_k]
|
75 |
+
|
76 |
+
all_data = collection.get(include=['metadatas'])['metadatas']
|
77 |
+
|
78 |
+
top_metadatas = [all_data[idx] for idx in top_indices]
|
79 |
+
|
80 |
results = []
|
81 |
+
for idx, metadata in enumerate(top_metadatas):
|
82 |
results.append({
|
83 |
+
'info': metadata,
|
84 |
+
'similarity': similarities[top_indices[idx]]
|
85 |
})
|
86 |
return results
|
87 |
|
88 |
+
|
89 |
+
|
90 |
def detect_clothing(image):
|
91 |
results = yolo_model(image)
|
92 |
detections = results[0].boxes.data.cpu().numpy()
|
|
|
163 |
cropped_image = crop_image(st.session_state.query_image, selected_detection['bbox'])
|
164 |
st.image(cropped_image, caption="Cropped Image", use_column_width=True)
|
165 |
query_embedding = get_image_embedding(cropped_image)
|
166 |
+
similar_images = find_similar_images(query_embedding, collection)
|
167 |
|
168 |
st.subheader("Similar Items:")
|
169 |
for img in similar_images:
|
|
|
189 |
if st.button("Search by Text"):
|
190 |
if query_text:
|
191 |
text_embedding = get_text_embedding(query_text)
|
192 |
+
similar_images = find_similar_images(text_embedding, collection)
|
193 |
st.subheader("Similar Items:")
|
194 |
for img in similar_images:
|
195 |
col1, col2 = st.columns(2)
|