Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
@@ -6,6 +6,7 @@ import numpy as np
|
|
6 |
import random
|
7 |
import os
|
8 |
import wget # To download weights
|
|
|
9 |
|
10 |
# --- Configuration & Model Loading ---
|
11 |
|
@@ -30,39 +31,68 @@ def load_clip_model():
|
|
30 |
print(f"CLIP model loaded to {DEVICE}.")
|
31 |
|
32 |
# --- FastSAM Setup ---
|
33 |
-
# Use a smaller model suitable for Spaces CPU/basic GPU if needed
|
34 |
FASTSAM_CHECKPOINT = "FastSAM-s.pt"
|
35 |
-
|
|
|
36 |
|
37 |
fastsam_model = None
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
38 |
|
39 |
def download_fastsam_weights():
|
40 |
if not os.path.exists(FASTSAM_CHECKPOINT):
|
41 |
-
print(f"Downloading FastSAM weights: {FASTSAM_CHECKPOINT}...")
|
42 |
try:
|
43 |
wget.download(FASTSAM_CHECKPOINT_URL, FASTSAM_CHECKPOINT)
|
44 |
print("FastSAM weights downloaded.")
|
45 |
except Exception as e:
|
46 |
print(f"Error downloading FastSAM weights: {e}")
|
47 |
print("Please ensure the URL is correct and reachable, or manually place the weights file.")
|
|
|
|
|
|
|
|
|
|
|
|
|
48 |
return False
|
49 |
return os.path.exists(FASTSAM_CHECKPOINT)
|
50 |
|
51 |
def load_fastsam_model():
|
52 |
global fastsam_model
|
53 |
if fastsam_model is None:
|
54 |
-
if
|
|
|
|
|
|
|
|
|
55 |
try:
|
56 |
-
|
57 |
print(f"Loading FastSAM model: {FASTSAM_CHECKPOINT}...")
|
58 |
fastsam_model = FastSAM(FASTSAM_CHECKPOINT)
|
59 |
-
print(f"FastSAM model loaded.") # Device handled internally by FastSAM
|
60 |
-
except ImportError:
|
61 |
-
print("Error: 'fastsam' library not found. Please install it (pip install fastsam).")
|
62 |
except Exception as e:
|
63 |
print(f"Error loading FastSAM model: {e}")
|
|
|
64 |
else:
|
65 |
-
print("FastSAM weights not found. Cannot load model.")
|
66 |
|
67 |
|
68 |
# --- Processing Functions ---
|
@@ -74,14 +104,16 @@ def run_clip_zero_shot(image: Image.Image, text_labels: str):
|
|
74 |
if clip_model is None:
|
75 |
return "Error: CLIP Model not loaded. Check logs.", None
|
76 |
|
77 |
-
if not text_labels:
|
78 |
-
return "Please provide comma-separated text labels.", None
|
79 |
if image is None:
|
80 |
-
return "Please upload an image.", None
|
|
|
|
|
|
|
81 |
|
82 |
-
labels = [label.strip() for label in text_labels.split(',')]
|
83 |
if not labels:
|
84 |
-
|
|
|
85 |
|
86 |
print(f"Running CLIP zero-shot classification with labels: {labels}")
|
87 |
|
@@ -94,28 +126,36 @@ def run_clip_zero_shot(image: Image.Image, text_labels: str):
|
|
94 |
|
95 |
with torch.no_grad():
|
96 |
outputs = clip_model(**inputs)
|
97 |
-
logits_per_image = outputs.logits_per_image
|
98 |
-
probs = logits_per_image.softmax(dim=1)
|
99 |
|
100 |
print("CLIP processing complete.")
|
101 |
|
102 |
-
# Format output for Gradio Label
|
103 |
confidences = {labels[i]: float(probs[0, i].item()) for i in range(len(labels))}
|
104 |
-
|
|
|
105 |
|
106 |
except Exception as e:
|
107 |
print(f"Error during CLIP processing: {e}")
|
108 |
-
|
|
|
|
|
109 |
|
110 |
|
111 |
# FastSAM Segmentation Function
|
112 |
def run_fastsam_segmentation(image_pil: Image.Image, conf_threshold: float = 0.4, iou_threshold: float = 0.9):
|
|
|
113 |
if fastsam_model is None:
|
114 |
-
load_fastsam_model()
|
115 |
if fastsam_model is None:
|
116 |
-
|
|
|
|
|
|
|
|
|
|
|
117 |
if image_pil is None:
|
118 |
-
return "Please upload an image."
|
119 |
|
120 |
print("Running FastSAM segmentation...")
|
121 |
|
@@ -124,63 +164,52 @@ def run_fastsam_segmentation(image_pil: Image.Image, conf_threshold: float = 0.4
|
|
124 |
if image_pil.mode != "RGB":
|
125 |
image_pil = image_pil.convert("RGB")
|
126 |
|
127 |
-
# FastSAM expects a BGR numpy array or path usually. Let's try with RGB numpy.
|
128 |
-
# If it fails, uncomment the BGR conversion line.
|
129 |
image_np_rgb = np.array(image_pil)
|
130 |
-
# image_np_bgr = image_np_rgb[:, :, ::-1] # Convert RGB to BGR if needed
|
131 |
|
132 |
# Run FastSAM inference
|
133 |
-
# Adjust imgsz, conf, iou as needed. Higher imgsz = more detail, slower.
|
134 |
everything_results = fastsam_model(
|
135 |
-
image_np_rgb,
|
136 |
device=DEVICE,
|
137 |
retina_masks=True,
|
138 |
-
imgsz=640,
|
139 |
conf=conf_threshold,
|
140 |
iou=iou_threshold,
|
141 |
)
|
142 |
|
143 |
-
#
|
144 |
-
from fastsam import FastSAMPrompt # Make sure it's imported
|
145 |
prompt_process = FastSAMPrompt(image_np_rgb, everything_results, device=DEVICE)
|
146 |
-
|
147 |
-
# Get all annotations (masks)
|
148 |
ann = prompt_process.everything_prompt()
|
149 |
|
150 |
-
print(f"FastSAM found {len(ann[0]['masks']) if ann and ann[0] else 0} masks.")
|
151 |
|
152 |
-
# --- Plotting Masks on Image
|
153 |
output_image = image_pil.copy()
|
154 |
if ann and ann[0] is not None and 'masks' in ann[0] and len(ann[0]['masks']) > 0:
|
155 |
-
masks = ann[0]['masks'].cpu().numpy() #
|
156 |
-
|
157 |
-
# Create overlay image
|
158 |
overlay = Image.new('RGBA', output_image.size, (0, 0, 0, 0))
|
159 |
draw = ImageDraw.Draw(overlay)
|
160 |
|
161 |
for i in range(masks.shape[0]):
|
162 |
-
mask = masks[i]
|
163 |
-
|
164 |
-
# Generate random color with some transparency
|
165 |
-
color = (random.randint(50, 255), random.randint(50, 255), random.randint(50, 255), 128) # RGBA with alpha
|
166 |
-
|
167 |
-
# Create a single-channel image from the boolean mask
|
168 |
mask_image = Image.fromarray((mask * 255).astype(np.uint8), mode='L')
|
169 |
-
|
170 |
-
# Apply color to the mask area on the overlay
|
171 |
draw.bitmap((0,0), mask_image, fill=color)
|
172 |
|
173 |
-
# Composite the overlay onto the original image
|
174 |
output_image = Image.alpha_composite(output_image.convert('RGBA'), overlay).convert('RGB')
|
175 |
|
176 |
print("FastSAM processing and plotting complete.")
|
177 |
-
|
|
|
178 |
|
|
|
|
|
|
|
|
|
179 |
except Exception as e:
|
180 |
print(f"Error during FastSAM processing: {e}")
|
181 |
-
|
182 |
-
|
183 |
-
return f"An error occurred: {e}", None
|
184 |
|
185 |
|
186 |
# --- Gradio Interface ---
|
@@ -188,7 +217,7 @@ def run_fastsam_segmentation(image_pil: Image.Image, conf_threshold: float = 0.4
|
|
188 |
# Pre-load models on startup (optional but good for performance)
|
189 |
print("Attempting to preload models...")
|
190 |
load_clip_model()
|
191 |
-
load_fastsam_model()
|
192 |
print("Preloading finished (or attempted).")
|
193 |
|
194 |
|
@@ -203,11 +232,11 @@ with gr.Blocks(theme=gr.themes.Soft()) as demo:
|
|
203 |
with gr.Row():
|
204 |
with gr.Column(scale=1):
|
205 |
clip_input_image = gr.Image(type="pil", label="Input Image")
|
206 |
-
clip_text_labels = gr.Textbox(label="Comma-Separated Labels", placeholder="e.g., astronaut,
|
207 |
clip_button = gr.Button("Run CLIP Classification", variant="primary")
|
208 |
with gr.Column(scale=1):
|
209 |
clip_output_label = gr.Label(label="Classification Probabilities")
|
210 |
-
clip_output_image_display = gr.Image(type="pil", label="Input Image Preview")
|
211 |
|
212 |
clip_button.click(
|
213 |
run_clip_zero_shot,
|
@@ -218,11 +247,12 @@ with gr.Blocks(theme=gr.themes.Soft()) as demo:
|
|
218 |
examples=[
|
219 |
["examples/astronaut.jpg", "astronaut, moon, rover, mountain"],
|
220 |
["examples/dog_bike.jpg", "dog, bicycle, person, park, grass"],
|
|
|
221 |
],
|
222 |
inputs=[clip_input_image, clip_text_labels],
|
223 |
outputs=[clip_output_label, clip_output_image_display],
|
224 |
fn=run_clip_zero_shot,
|
225 |
-
cache_examples=False,
|
226 |
)
|
227 |
|
228 |
# --- FastSAM Tab ---
|
@@ -237,41 +267,51 @@ with gr.Blocks(theme=gr.themes.Soft()) as demo:
|
|
237 |
fastsam_button = gr.Button("Run FastSAM Segmentation", variant="primary")
|
238 |
with gr.Column(scale=1):
|
239 |
fastsam_output_image = gr.Image(type="pil", label="Segmented Image")
|
240 |
-
# fastsam_input_display = gr.Image(type="pil", label="Original Image") # Optional: show original side-by-side
|
241 |
|
242 |
fastsam_button.click(
|
243 |
run_fastsam_segmentation,
|
244 |
inputs=[fastsam_input_image, fastsam_conf, fastsam_iou],
|
245 |
-
|
|
|
246 |
)
|
247 |
gr.Examples(
|
248 |
examples=[
|
249 |
["examples/dogs.jpg", 0.4, 0.9],
|
250 |
["examples/fruits.jpg", 0.5, 0.8],
|
|
|
251 |
],
|
252 |
inputs=[fastsam_input_image, fastsam_conf, fastsam_iou],
|
253 |
outputs=[fastsam_output_image],
|
254 |
fn=run_fastsam_segmentation,
|
255 |
-
cache_examples=False,
|
256 |
)
|
257 |
|
258 |
# Add example images (optional, but helpful)
|
259 |
-
# Create an 'examples' folder and add some jpg images like 'astronaut.jpg', 'dog_bike.jpg', 'dogs.jpg', 'fruits.jpg'
|
260 |
if not os.path.exists("examples"):
|
261 |
os.makedirs("examples")
|
262 |
-
print("Created 'examples' directory.
|
263 |
-
|
264 |
-
|
265 |
-
|
266 |
-
|
267 |
-
|
268 |
-
|
269 |
-
|
270 |
-
|
271 |
-
|
272 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
273 |
|
274 |
|
275 |
# Launch the Gradio app
|
276 |
if __name__ == "__main__":
|
277 |
-
|
|
|
|
|
|
|
|
6 |
import random
|
7 |
import os
|
8 |
import wget # To download weights
|
9 |
+
import traceback # For detailed error printing
|
10 |
|
11 |
# --- Configuration & Model Loading ---
|
12 |
|
|
|
31 |
print(f"CLIP model loaded to {DEVICE}.")
|
32 |
|
33 |
# --- FastSAM Setup ---
|
|
|
34 |
FASTSAM_CHECKPOINT = "FastSAM-s.pt"
|
35 |
+
# Use the official model hub repo URL
|
36 |
+
FASTSAM_CHECKPOINT_URL = f"https://huggingface.co/CASIA-IVA-Lab/FastSAM-s/resolve/main/{FASTSAM_CHECKPOINT}"
|
37 |
|
38 |
fastsam_model = None
|
39 |
+
fastsam_lib_imported = False # Flag to check if import worked
|
40 |
+
|
41 |
+
def check_and_import_fastsam():
|
42 |
+
global fastsam_lib_imported
|
43 |
+
if not fastsam_lib_imported:
|
44 |
+
try:
|
45 |
+
from fastsam import FastSAM, FastSAMPrompt
|
46 |
+
globals()['FastSAM'] = FastSAM # Make classes available globally
|
47 |
+
globals()['FastSAMPrompt'] = FastSAMPrompt
|
48 |
+
fastsam_lib_imported = True
|
49 |
+
print("fastsam library imported successfully.")
|
50 |
+
except ImportError:
|
51 |
+
print("Error: 'fastsam' library not found or import failed.")
|
52 |
+
print("Please ensure 'fastsam' is installed correctly (pip install fastsam).")
|
53 |
+
fastsam_lib_imported = False
|
54 |
+
except Exception as e:
|
55 |
+
print(f"An unexpected error occurred during fastsam import: {e}")
|
56 |
+
fastsam_lib_imported = False
|
57 |
+
return fastsam_lib_imported
|
58 |
+
|
59 |
|
60 |
def download_fastsam_weights():
|
61 |
if not os.path.exists(FASTSAM_CHECKPOINT):
|
62 |
+
print(f"Downloading FastSAM weights: {FASTSAM_CHECKPOINT} from {FASTSAM_CHECKPOINT_URL}...")
|
63 |
try:
|
64 |
wget.download(FASTSAM_CHECKPOINT_URL, FASTSAM_CHECKPOINT)
|
65 |
print("FastSAM weights downloaded.")
|
66 |
except Exception as e:
|
67 |
print(f"Error downloading FastSAM weights: {e}")
|
68 |
print("Please ensure the URL is correct and reachable, or manually place the weights file.")
|
69 |
+
# Attempt to remove partially downloaded file if exists
|
70 |
+
if os.path.exists(FASTSAM_CHECKPOINT):
|
71 |
+
try:
|
72 |
+
os.remove(FASTSAM_CHECKPOINT)
|
73 |
+
except OSError:
|
74 |
+
pass # Ignore removal errors
|
75 |
return False
|
76 |
return os.path.exists(FASTSAM_CHECKPOINT)
|
77 |
|
78 |
def load_fastsam_model():
|
79 |
global fastsam_model
|
80 |
if fastsam_model is None:
|
81 |
+
if not check_and_import_fastsam(): # Check import first
|
82 |
+
print("Cannot load FastSAM model because the library couldn't be imported.")
|
83 |
+
return # Exit if import failed
|
84 |
+
|
85 |
+
if download_fastsam_weights(): # Check download/existence second
|
86 |
try:
|
87 |
+
# FastSAM class should be available via globals() now
|
88 |
print(f"Loading FastSAM model: {FASTSAM_CHECKPOINT}...")
|
89 |
fastsam_model = FastSAM(FASTSAM_CHECKPOINT)
|
90 |
+
print(f"FastSAM model loaded.") # Device handled internally by FastSAM
|
|
|
|
|
91 |
except Exception as e:
|
92 |
print(f"Error loading FastSAM model: {e}")
|
93 |
+
traceback.print_exc()
|
94 |
else:
|
95 |
+
print("FastSAM weights not found or download failed. Cannot load model.")
|
96 |
|
97 |
|
98 |
# --- Processing Functions ---
|
|
|
104 |
if clip_model is None:
|
105 |
return "Error: CLIP Model not loaded. Check logs.", None
|
106 |
|
|
|
|
|
107 |
if image is None:
|
108 |
+
return "Please upload an image.", None # Return None for the image display
|
109 |
+
if not text_labels:
|
110 |
+
# Return empty results but display the uploaded image
|
111 |
+
return {}, image
|
112 |
|
113 |
+
labels = [label.strip() for label in text_labels.split(',') if label.strip()] # Ensure non-empty labels
|
114 |
if not labels:
|
115 |
+
# Return empty results but display the uploaded image
|
116 |
+
return {}, image
|
117 |
|
118 |
print(f"Running CLIP zero-shot classification with labels: {labels}")
|
119 |
|
|
|
126 |
|
127 |
with torch.no_grad():
|
128 |
outputs = clip_model(**inputs)
|
129 |
+
logits_per_image = outputs.logits_per_image
|
130 |
+
probs = logits_per_image.softmax(dim=1)
|
131 |
|
132 |
print("CLIP processing complete.")
|
133 |
|
|
|
134 |
confidences = {labels[i]: float(probs[0, i].item()) for i in range(len(labels))}
|
135 |
+
# Return results and the original image used for prediction
|
136 |
+
return confidences, image
|
137 |
|
138 |
except Exception as e:
|
139 |
print(f"Error during CLIP processing: {e}")
|
140 |
+
traceback.print_exc()
|
141 |
+
# Return error message and the original image
|
142 |
+
return f"An error occurred during CLIP: {e}", image
|
143 |
|
144 |
|
145 |
# FastSAM Segmentation Function
|
146 |
def run_fastsam_segmentation(image_pil: Image.Image, conf_threshold: float = 0.4, iou_threshold: float = 0.9):
|
147 |
+
# Ensure model is loaded or attempt to load
|
148 |
if fastsam_model is None:
|
149 |
+
load_fastsam_model()
|
150 |
if fastsam_model is None:
|
151 |
+
# Return error message string for the image component (Gradio handles this)
|
152 |
+
return "Error: FastSAM Model not loaded. Check logs."
|
153 |
+
# Ensure library was imported
|
154 |
+
if not fastsam_lib_imported:
|
155 |
+
return "Error: FastSAM library not available. Cannot run segmentation."
|
156 |
+
|
157 |
if image_pil is None:
|
158 |
+
return "Please upload an image."
|
159 |
|
160 |
print("Running FastSAM segmentation...")
|
161 |
|
|
|
164 |
if image_pil.mode != "RGB":
|
165 |
image_pil = image_pil.convert("RGB")
|
166 |
|
|
|
|
|
167 |
image_np_rgb = np.array(image_pil)
|
|
|
168 |
|
169 |
# Run FastSAM inference
|
|
|
170 |
everything_results = fastsam_model(
|
171 |
+
image_np_rgb,
|
172 |
device=DEVICE,
|
173 |
retina_masks=True,
|
174 |
+
imgsz=640,
|
175 |
conf=conf_threshold,
|
176 |
iou=iou_threshold,
|
177 |
)
|
178 |
|
179 |
+
# FastSAMPrompt should be available via globals() if import succeeded
|
|
|
180 |
prompt_process = FastSAMPrompt(image_np_rgb, everything_results, device=DEVICE)
|
|
|
|
|
181 |
ann = prompt_process.everything_prompt()
|
182 |
|
183 |
+
print(f"FastSAM found {len(ann[0]['masks']) if ann and ann[0] and 'masks' in ann[0] else 0} masks.")
|
184 |
|
185 |
+
# --- Plotting Masks on Image ---
|
186 |
output_image = image_pil.copy()
|
187 |
if ann and ann[0] is not None and 'masks' in ann[0] and len(ann[0]['masks']) > 0:
|
188 |
+
masks = ann[0]['masks'].cpu().numpy() # (N, H, W) boolean
|
189 |
+
|
|
|
190 |
overlay = Image.new('RGBA', output_image.size, (0, 0, 0, 0))
|
191 |
draw = ImageDraw.Draw(overlay)
|
192 |
|
193 |
for i in range(masks.shape[0]):
|
194 |
+
mask = masks[i]
|
195 |
+
color = (random.randint(50, 255), random.randint(50, 255), random.randint(50, 255), 128) # RGBA
|
|
|
|
|
|
|
|
|
196 |
mask_image = Image.fromarray((mask * 255).astype(np.uint8), mode='L')
|
|
|
|
|
197 |
draw.bitmap((0,0), mask_image, fill=color)
|
198 |
|
|
|
199 |
output_image = Image.alpha_composite(output_image.convert('RGBA'), overlay).convert('RGB')
|
200 |
|
201 |
print("FastSAM processing and plotting complete.")
|
202 |
+
# *** FIX: Return ONLY the output image for the single Image component ***
|
203 |
+
return output_image
|
204 |
|
205 |
+
except NameError as ne:
|
206 |
+
print(f"NameError during FastSAM processing: {ne}. Was the fastsam library imported correctly?")
|
207 |
+
traceback.print_exc()
|
208 |
+
return f"A NameError occurred: {ne}. Check library import."
|
209 |
except Exception as e:
|
210 |
print(f"Error during FastSAM processing: {e}")
|
211 |
+
traceback.print_exc()
|
212 |
+
return f"An error occurred during FastSAM: {e}"
|
|
|
213 |
|
214 |
|
215 |
# --- Gradio Interface ---
|
|
|
217 |
# Pre-load models on startup (optional but good for performance)
|
218 |
print("Attempting to preload models...")
|
219 |
load_clip_model()
|
220 |
+
load_fastsam_model() # This will now also attempt download/check import
|
221 |
print("Preloading finished (or attempted).")
|
222 |
|
223 |
|
|
|
232 |
with gr.Row():
|
233 |
with gr.Column(scale=1):
|
234 |
clip_input_image = gr.Image(type="pil", label="Input Image")
|
235 |
+
clip_text_labels = gr.Textbox(label="Comma-Separated Labels", placeholder="e.g., astronaut, moon, dog playing fetch")
|
236 |
clip_button = gr.Button("Run CLIP Classification", variant="primary")
|
237 |
with gr.Column(scale=1):
|
238 |
clip_output_label = gr.Label(label="Classification Probabilities")
|
239 |
+
clip_output_image_display = gr.Image(type="pil", label="Input Image Preview")
|
240 |
|
241 |
clip_button.click(
|
242 |
run_clip_zero_shot,
|
|
|
247 |
examples=[
|
248 |
["examples/astronaut.jpg", "astronaut, moon, rover, mountain"],
|
249 |
["examples/dog_bike.jpg", "dog, bicycle, person, park, grass"],
|
250 |
+
["examples/clip_logo.png", "logo, text, graphics, abstract art"], # Added another example
|
251 |
],
|
252 |
inputs=[clip_input_image, clip_text_labels],
|
253 |
outputs=[clip_output_label, clip_output_image_display],
|
254 |
fn=run_clip_zero_shot,
|
255 |
+
cache_examples=False,
|
256 |
)
|
257 |
|
258 |
# --- FastSAM Tab ---
|
|
|
267 |
fastsam_button = gr.Button("Run FastSAM Segmentation", variant="primary")
|
268 |
with gr.Column(scale=1):
|
269 |
fastsam_output_image = gr.Image(type="pil", label="Segmented Image")
|
|
|
270 |
|
271 |
fastsam_button.click(
|
272 |
run_fastsam_segmentation,
|
273 |
inputs=[fastsam_input_image, fastsam_conf, fastsam_iou],
|
274 |
+
# Output is now correctly mapped to the single component
|
275 |
+
outputs=[fastsam_output_image]
|
276 |
)
|
277 |
gr.Examples(
|
278 |
examples=[
|
279 |
["examples/dogs.jpg", 0.4, 0.9],
|
280 |
["examples/fruits.jpg", 0.5, 0.8],
|
281 |
+
["examples/lion.jpg", 0.45, 0.9], # Added another example
|
282 |
],
|
283 |
inputs=[fastsam_input_image, fastsam_conf, fastsam_iou],
|
284 |
outputs=[fastsam_output_image],
|
285 |
fn=run_fastsam_segmentation,
|
286 |
+
cache_examples=False,
|
287 |
)
|
288 |
|
289 |
# Add example images (optional, but helpful)
|
|
|
290 |
if not os.path.exists("examples"):
|
291 |
os.makedirs("examples")
|
292 |
+
print("Created 'examples' directory. Attempting to download sample images...")
|
293 |
+
example_files = {
|
294 |
+
"astronaut.jpg": "https://upload.wikimedia.org/wikipedia/commons/thumb/d/d astronaut_-_St._Jean_Bay.jpg/640px-Astronaut_-_St._Jean_Bay.jpg", # Find suitable public domain/CC image
|
295 |
+
"dog_bike.jpg": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/gradio/outputs_multimodal.jpg", # Using a relevant example from HF
|
296 |
+
"clip_logo.png": "https://raw.githubusercontent.com/openai/CLIP/main/CLIP.png",
|
297 |
+
"dogs.jpg": "https://raw.githubusercontent.com/ultralytics/assets/main/im/image8.jpg", # From Ultralytics assets
|
298 |
+
"fruits.jpg": "https://raw.githubusercontent.com/ultralytics/assets/main/im/image9.jpg", # From Ultralytics assets
|
299 |
+
"lion.jpg": "https://huggingface.co/spaces/gradio/image-segmentation/resolve/main/images/lion.jpg"
|
300 |
+
}
|
301 |
+
for filename, url in example_files.items():
|
302 |
+
filepath = os.path.join("examples", filename)
|
303 |
+
if not os.path.exists(filepath):
|
304 |
+
try:
|
305 |
+
print(f"Downloading {filename}...")
|
306 |
+
wget.download(url, filepath)
|
307 |
+
except Exception as e:
|
308 |
+
print(f"Could not download {filename} from {url}: {e}")
|
309 |
+
print("Example image download attempt finished.")
|
310 |
|
311 |
|
312 |
# Launch the Gradio app
|
313 |
if __name__ == "__main__":
|
314 |
+
# share=True is primarily for local testing to get a public link.
|
315 |
+
# Not needed/used when deploying on Hugging Face Spaces.
|
316 |
+
# debug=True is helpful for development. Set to False for production.
|
317 |
+
demo.launch(debug=True)
|