Spaces:
Build error
Build error
RobotJelly
commited on
Commit
·
7c286b0
1
Parent(s):
4a33ca4
app.py
Browse files
app.py
CHANGED
@@ -3,25 +3,27 @@ from pathlib import Path
|
|
3 |
import pandas as pd
|
4 |
import numpy as np
|
5 |
import torch
|
|
|
6 |
from PIL import Image
|
7 |
from io import BytesIO
|
8 |
import requests
|
9 |
import gradio as gr
|
10 |
import os
|
11 |
from transformers import CLIPProcessor, CLIPModel, CLIPTokenizer
|
12 |
-
import
|
|
|
13 |
|
14 |
# check if CUDA available
|
15 |
device = "cuda" if torch.cuda.is_available() else "cpu"
|
16 |
|
17 |
# Load the openAI's CLIP model
|
18 |
-
model = CLIPModel.from_pretrained("openai/clip-vit-base-patch32")
|
19 |
-
processor = CLIPProcessor.from_pretrained("openai/clip-vit-base-patch32")
|
20 |
-
tokenizer = CLIPTokenizer.from_pretrained("openai/clip-vit-base-patch32")
|
21 |
|
22 |
# taking photo IDs
|
23 |
-
photo_ids = pd.read_csv("./photo_ids.csv")
|
24 |
-
photo_ids = list(photo_ids['photo_id'])
|
25 |
|
26 |
# Photo dataset
|
27 |
photos = pd.read_csv("./photos.tsv000", sep="\t", header=0)
|
@@ -31,38 +33,56 @@ photo_features = np.load("./features.npy")
|
|
31 |
|
32 |
IMAGES_DIR = './photos'
|
33 |
|
34 |
-
|
35 |
-
|
36 |
-
|
37 |
-
|
|
|
|
|
38 |
#response = requests.get(photo_image_url, stream=True)
|
39 |
#img = Image.open(BytesIO(response.content))
|
40 |
-
|
41 |
-
|
42 |
#photo = photo_id + '.jpg'
|
43 |
#img = Image.open(response).convert("RGB")
|
44 |
#img = Image.open(os.path.join(IMAGES_DIR, photo))
|
45 |
-
image.append(img)
|
46 |
-
return image
|
|
|
47 |
|
48 |
# Encode and normalize the search query using CLIP
|
49 |
-
def encode_search_query(search_query, model, device):
|
50 |
-
with torch.no_grad():
|
51 |
-
inputs = tokenizer([search_query], padding=True, return_tensors="pt")
|
52 |
#inputs = processor(text=[search_query], images=None, return_tensors="pt", padding=True)
|
53 |
-
text_features = model.get_text_features(**inputs).cpu().numpy()
|
54 |
-
return text_features
|
55 |
|
56 |
# Find all matched photos
|
57 |
-
def find_matches(features, photo_ids, results_count=4):
|
58 |
# Compute the similarity between the search query and each photo using the Cosine similarity
|
59 |
#text_features = np.array(text_features)
|
60 |
-
similarities = (photo_features @ features.T).squeeze(1)
|
61 |
# Sort the photos by their similarity score
|
62 |
-
best_photo_idx = (-similarities).argsort()
|
63 |
# Return the photo IDs of the best matches
|
64 |
-
matches = [photo_ids[i] for i in best_photo_idx[:results_count]]
|
65 |
-
return matches
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
66 |
|
67 |
def image_search(search_text, search_image, option):
|
68 |
|
@@ -70,25 +90,35 @@ def image_search(search_text, search_image, option):
|
|
70 |
#search_query = "The feeling when your program finally works"
|
71 |
|
72 |
if option == "Text-To-Image" :
|
73 |
-
# Extracting text features
|
74 |
-
text_features = encode_search_query(search_text, model, device)
|
|
|
75 |
|
76 |
# Find the matched Images
|
77 |
-
matched_images = find_matches(text_features, photo_features, photo_ids, 4)
|
|
|
78 |
|
79 |
-
|
|
|
80 |
elif option == "Image-To-Image":
|
81 |
# Input Image for Search
|
82 |
#search_image = Image.fromarray(search_image.astype('uint8'), 'RGB')
|
83 |
|
84 |
-
with torch.no_grad():
|
85 |
-
|
86 |
-
|
87 |
-
|
88 |
-
image_feature = image_feature.cpu().numpy()
|
89 |
# Find the matched Images
|
90 |
-
matched_images = find_matches(image_feature, photo_ids, 4)
|
91 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
92 |
|
93 |
gr.Interface(fn=image_search,
|
94 |
inputs=[gr.inputs.Textbox(lines=7, label="Input Text"),
|
|
|
3 |
import pandas as pd
|
4 |
import numpy as np
|
5 |
import torch
|
6 |
+
import pickle
|
7 |
from PIL import Image
|
8 |
from io import BytesIO
|
9 |
import requests
|
10 |
import gradio as gr
|
11 |
import os
|
12 |
from transformers import CLIPProcessor, CLIPModel, CLIPTokenizer
|
13 |
+
import sentence_transformers
|
14 |
+
from sentence_transformers import SentenceTransformer, util
|
15 |
|
16 |
# check if CUDA available
|
17 |
device = "cuda" if torch.cuda.is_available() else "cpu"
|
18 |
|
19 |
# Load the openAI's CLIP model
|
20 |
+
#model = CLIPModel.from_pretrained("openai/clip-vit-base-patch32")
|
21 |
+
#processor = CLIPProcessor.from_pretrained("openai/clip-vit-base-patch32")
|
22 |
+
#tokenizer = CLIPTokenizer.from_pretrained("openai/clip-vit-base-patch32")
|
23 |
|
24 |
# taking photo IDs
|
25 |
+
#photo_ids = pd.read_csv("./photo_ids.csv")
|
26 |
+
#photo_ids = list(photo_ids['photo_id'])
|
27 |
|
28 |
# Photo dataset
|
29 |
photos = pd.read_csv("./photos.tsv000", sep="\t", header=0)
|
|
|
33 |
|
34 |
IMAGES_DIR = './photos'
|
35 |
|
36 |
+
|
37 |
+
|
38 |
+
#def show_output_image(matched_images) :
|
39 |
+
#image=[]
|
40 |
+
#for photo_id in matched_images:
|
41 |
+
# photo_image_url = f"https://unsplash.com/photos/{photo_id}/download?w=280"
|
42 |
#response = requests.get(photo_image_url, stream=True)
|
43 |
#img = Image.open(BytesIO(response.content))
|
44 |
+
# response = requests.get(photo_image_url, stream=True).raw
|
45 |
+
# img = Image.open(response)
|
46 |
#photo = photo_id + '.jpg'
|
47 |
#img = Image.open(response).convert("RGB")
|
48 |
#img = Image.open(os.path.join(IMAGES_DIR, photo))
|
49 |
+
#image.append(img)
|
50 |
+
#return image
|
51 |
+
|
52 |
|
53 |
# Encode and normalize the search query using CLIP
|
54 |
+
#def encode_search_query(search_query, model, device):
|
55 |
+
# with torch.no_grad():
|
56 |
+
# inputs = tokenizer([search_query], padding=True, return_tensors="pt")
|
57 |
#inputs = processor(text=[search_query], images=None, return_tensors="pt", padding=True)
|
58 |
+
# text_features = model.get_text_features(**inputs).cpu().numpy()
|
59 |
+
# return text_features
|
60 |
|
61 |
# Find all matched photos
|
62 |
+
#def find_matches(features, photo_ids, results_count=4):
|
63 |
# Compute the similarity between the search query and each photo using the Cosine similarity
|
64 |
#text_features = np.array(text_features)
|
65 |
+
#similarities = (photo_features @ features.T).squeeze(1)
|
66 |
# Sort the photos by their similarity score
|
67 |
+
#best_photo_idx = (-similarities).argsort()
|
68 |
# Return the photo IDs of the best matches
|
69 |
+
#matches = [photo_ids[i] for i in best_photo_idx[:results_count]]
|
70 |
+
#return matches
|
71 |
+
|
72 |
+
#Load CLIP model
|
73 |
+
model = SentenceTransformer('clip-ViT-B-32')
|
74 |
+
|
75 |
+
# pre-computed embeddings
|
76 |
+
emb_filename = 'unsplash-25k-photos-embeddings.pkl'
|
77 |
+
with open(emb_filename, 'rb') as fIn:
|
78 |
+
img_names, img_emb = pickle.load(fIn)
|
79 |
+
|
80 |
+
def display_matches(similarity):
|
81 |
+
best_matched_images = []
|
82 |
+
for best_img in torch.topk(similarity, 4, 0).indices:
|
83 |
+
img = Image.open(os.path.join('./photos', img_names[best_img]))
|
84 |
+
best_matched_images.append(img)
|
85 |
+
return best_matched_images
|
86 |
|
87 |
def image_search(search_text, search_image, option):
|
88 |
|
|
|
90 |
#search_query = "The feeling when your program finally works"
|
91 |
|
92 |
if option == "Text-To-Image" :
|
93 |
+
# Extracting text features embeddings
|
94 |
+
#text_features = encode_search_query(search_text, model, device)
|
95 |
+
text_emb = model.encode([serach_text], convert_to_tensor=True)
|
96 |
|
97 |
# Find the matched Images
|
98 |
+
#matched_images = find_matches(text_features, photo_features, photo_ids, 4)
|
99 |
+
similarity = util.cos_sim(text_emb, img_emb)
|
100 |
|
101 |
+
# top 4 highest ranked images
|
102 |
+
return display_matches(similarity)
|
103 |
elif option == "Image-To-Image":
|
104 |
# Input Image for Search
|
105 |
#search_image = Image.fromarray(search_image.astype('uint8'), 'RGB')
|
106 |
|
107 |
+
#with torch.no_grad():
|
108 |
+
# processed_image = processor(text=None, images=search_image, return_tensors="pt", padding=True)["pixel_values"]
|
109 |
+
# image_feature = model.get_image_features(processed_image.to(device))
|
110 |
+
# image_feature /= image_feature.norm(dim=-1, keepdim=True)
|
111 |
+
#image_feature = image_feature.cpu().numpy()
|
112 |
# Find the matched Images
|
113 |
+
#matched_images = find_matches(image_feature, photo_ids, 4)
|
114 |
+
|
115 |
+
image_emb = model.encode(Image.open(search_image), convert_to_tensor=True)
|
116 |
+
|
117 |
+
# Find the matched Images
|
118 |
+
#matched_images = find_matches(text_features, photo_features, photo_ids, 4)
|
119 |
+
similarity = util.cos_sim(image_emb, img_emb)
|
120 |
+
|
121 |
+
return display_matches(similarity)
|
122 |
|
123 |
gr.Interface(fn=image_search,
|
124 |
inputs=[gr.inputs.Textbox(lines=7, label="Input Text"),
|