Spaces:
Running
on
Zero
Running
on
Zero
RageshAntony
commited on
new flow
Browse files
app.py
CHANGED
@@ -15,12 +15,17 @@ import threading
|
|
15 |
from pathlib import Path
|
16 |
import shutil
|
17 |
import time
|
|
|
|
|
|
|
18 |
|
19 |
# Constants
|
20 |
MAX_SEED = np.iinfo(np.int32).max
|
21 |
MAX_IMAGE_SIZE = 1024
|
22 |
DEVICE = "cuda" if torch.cuda.is_available() else "cpu"
|
23 |
TORCH_DTYPE = torch.bfloat16 if torch.cuda.is_available() else torch.float32
|
|
|
|
|
24 |
|
25 |
# Model configurations
|
26 |
MODEL_CONFIGS = {
|
@@ -135,7 +140,30 @@ def load_pipeline(model_name):
|
|
135 |
|
136 |
return pipe
|
137 |
|
138 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
139 |
def generate_image(
|
140 |
model_name,
|
141 |
prompt,
|
@@ -152,7 +180,6 @@ def generate_image(
|
|
152 |
try:
|
153 |
progress(0, desc=f"Loading {model_name} model...")
|
154 |
|
155 |
-
# Load model if not already loaded
|
156 |
if model_name not in pipes:
|
157 |
pipes[model_name] = load_pipeline(model_name)
|
158 |
|
@@ -165,7 +192,6 @@ def generate_image(
|
|
165 |
print(f"Generating image with {model_name}...")
|
166 |
progress(0.3, desc=f"Generating image with {model_name}...")
|
167 |
|
168 |
-
# Generate image
|
169 |
image = pipe(
|
170 |
prompt=prompt,
|
171 |
negative_prompt=negative_prompt,
|
@@ -176,9 +202,10 @@ def generate_image(
|
|
176 |
generator=generator,
|
177 |
).images[0]
|
178 |
|
179 |
-
|
|
|
180 |
|
181 |
-
|
182 |
deep_cleanup(model_name, pipe)
|
183 |
|
184 |
progress(1.0, desc=f"Generation complete with {model_name}")
|
@@ -186,7 +213,6 @@ def generate_image(
|
|
186 |
|
187 |
except Exception as e:
|
188 |
print(f"Error with {model_name}: {str(e)}")
|
189 |
-
# Ensure cleanup happens even if generation fails
|
190 |
if model_name in pipes:
|
191 |
deep_cleanup(model_name, pipes[model_name])
|
192 |
raise e
|
@@ -198,7 +224,6 @@ css = """
|
|
198 |
max-width: 1024px;
|
199 |
}
|
200 |
"""
|
201 |
-
#run_test_safe.zerogpu = True
|
202 |
|
203 |
with gr.Blocks(css=css) as demo:
|
204 |
with gr.Column(elem_id="col-container"):
|
@@ -263,17 +288,28 @@ with gr.Blocks(css=css) as demo:
|
|
263 |
value=40,
|
264 |
)
|
265 |
|
266 |
-
# Memory usage indicator
|
267 |
memory_indicator = gr.Markdown("Current memory usage: 0 GB")
|
268 |
|
269 |
-
|
270 |
-
|
271 |
-
|
272 |
-
|
273 |
-
|
274 |
-
|
275 |
-
|
276 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
277 |
|
278 |
examples = [
|
279 |
"A capybara wearing a suit holding a sign that reads Hello World",
|
@@ -281,15 +317,14 @@ with gr.Blocks(css=css) as demo:
|
|
281 |
]
|
282 |
gr.Examples(examples=examples, inputs=[prompt])
|
283 |
|
284 |
-
def
|
285 |
-
"""Update
|
286 |
-
|
287 |
-
|
288 |
-
|
289 |
-
|
290 |
-
|
291 |
|
292 |
-
# Handle generation for each model
|
293 |
@spaces.GPU(duration=600)
|
294 |
def generate_all(prompt, negative_prompt, seed, randomize_seed, width, height, guidance_scale, num_inference_steps, progress=gr.Progress()):
|
295 |
outputs = []
|
@@ -304,16 +339,16 @@ with gr.Blocks(css=css) as demo:
|
|
304 |
print(f"IMAGE GENERATED {model_name} {update_memory_usage()}")
|
305 |
outputs.extend([image, used_seed])
|
306 |
|
307 |
-
# Update memory usage after each model
|
308 |
-
#memory_indicator.update(update_memory_usage())
|
309 |
-
|
310 |
except Exception as e:
|
311 |
outputs.extend([None, None])
|
312 |
print(f"Error generating with {model_name}: {str(e)}")
|
313 |
|
|
|
|
|
|
|
|
|
314 |
return outputs
|
315 |
|
316 |
-
# Set up the generation trigger
|
317 |
output_components = []
|
318 |
for model_name in MODEL_CONFIGS.keys():
|
319 |
output_components.extend([results[model_name], seeds[model_name]])
|
@@ -333,5 +368,17 @@ with gr.Blocks(css=css) as demo:
|
|
333 |
outputs=output_components,
|
334 |
)
|
335 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
336 |
if __name__ == "__main__":
|
337 |
demo.launch()
|
|
|
15 |
from pathlib import Path
|
16 |
import shutil
|
17 |
import time
|
18 |
+
import glob
|
19 |
+
from datetime import datetime
|
20 |
+
from PIL import Image
|
21 |
|
22 |
# Constants
|
23 |
MAX_SEED = np.iinfo(np.int32).max
|
24 |
MAX_IMAGE_SIZE = 1024
|
25 |
DEVICE = "cuda" if torch.cuda.is_available() else "cpu"
|
26 |
TORCH_DTYPE = torch.bfloat16 if torch.cuda.is_available() else torch.float32
|
27 |
+
OUTPUT_DIR = "generated_images"
|
28 |
+
os.makedirs(OUTPUT_DIR, exist_ok=True)
|
29 |
|
30 |
# Model configurations
|
31 |
MODEL_CONFIGS = {
|
|
|
140 |
|
141 |
return pipe
|
142 |
|
143 |
+
def save_generated_image(image, model_name, prompt):
|
144 |
+
"""Save generated image with timestamp and model name"""
|
145 |
+
timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
|
146 |
+
# Create sanitized filename from prompt (first 30 chars)
|
147 |
+
prompt_part = "".join(c for c in prompt[:30] if c.isalnum() or c in (' ', '-', '_')).strip()
|
148 |
+
filename = f"{timestamp}_{model_name}_{prompt_part}.png"
|
149 |
+
filepath = os.path.join(OUTPUT_DIR, filename)
|
150 |
+
image.save(filepath)
|
151 |
+
return filepath
|
152 |
+
|
153 |
+
def get_generated_images():
|
154 |
+
"""Get list of generated images with their details"""
|
155 |
+
files = glob.glob(os.path.join(OUTPUT_DIR, "*.png"))
|
156 |
+
files.sort(key=os.path.getctime, reverse=True) # Sort by creation time
|
157 |
+
return [
|
158 |
+
{
|
159 |
+
"path": f,
|
160 |
+
"name": os.path.basename(f),
|
161 |
+
"date": datetime.fromtimestamp(os.path.getctime(f)).strftime("%Y-%m-%d %H:%M:%S"),
|
162 |
+
"size": f"{os.path.getsize(f) / 1024:.1f} KB"
|
163 |
+
}
|
164 |
+
for f in files
|
165 |
+
]
|
166 |
+
|
167 |
def generate_image(
|
168 |
model_name,
|
169 |
prompt,
|
|
|
180 |
try:
|
181 |
progress(0, desc=f"Loading {model_name} model...")
|
182 |
|
|
|
183 |
if model_name not in pipes:
|
184 |
pipes[model_name] = load_pipeline(model_name)
|
185 |
|
|
|
192 |
print(f"Generating image with {model_name}...")
|
193 |
progress(0.3, desc=f"Generating image with {model_name}...")
|
194 |
|
|
|
195 |
image = pipe(
|
196 |
prompt=prompt,
|
197 |
negative_prompt=negative_prompt,
|
|
|
202 |
generator=generator,
|
203 |
).images[0]
|
204 |
|
205 |
+
filepath = save_generated_image(image, model_name, prompt)
|
206 |
+
print(f"Saved image to: {filepath}")
|
207 |
|
208 |
+
progress(0.9, desc=f"Cleaning up {model_name} resources...")
|
209 |
deep_cleanup(model_name, pipe)
|
210 |
|
211 |
progress(1.0, desc=f"Generation complete with {model_name}")
|
|
|
213 |
|
214 |
except Exception as e:
|
215 |
print(f"Error with {model_name}: {str(e)}")
|
|
|
216 |
if model_name in pipes:
|
217 |
deep_cleanup(model_name, pipes[model_name])
|
218 |
raise e
|
|
|
224 |
max-width: 1024px;
|
225 |
}
|
226 |
"""
|
|
|
227 |
|
228 |
with gr.Blocks(css=css) as demo:
|
229 |
with gr.Column(elem_id="col-container"):
|
|
|
288 |
value=40,
|
289 |
)
|
290 |
|
|
|
291 |
memory_indicator = gr.Markdown("Current memory usage: 0 GB")
|
292 |
|
293 |
+
with gr.Row():
|
294 |
+
with gr.Column(scale=2):
|
295 |
+
with gr.Tabs() as tabs:
|
296 |
+
results = {}
|
297 |
+
seeds = {}
|
298 |
+
for model_name in MODEL_CONFIGS.keys():
|
299 |
+
with gr.Tab(model_name):
|
300 |
+
results[model_name] = gr.Image(label=f"{model_name} Result")
|
301 |
+
seeds[model_name] = gr.Number(label="Seed used", visible=False)
|
302 |
+
|
303 |
+
with gr.Column(scale=1):
|
304 |
+
gr.Markdown("### Generated Images")
|
305 |
+
file_gallery = gr.Gallery(
|
306 |
+
label="Generated Images",
|
307 |
+
show_label=False,
|
308 |
+
elem_id="file_gallery",
|
309 |
+
columns=2,
|
310 |
+
height=400
|
311 |
+
)
|
312 |
+
refresh_button = gr.Button("Refresh Gallery")
|
313 |
|
314 |
examples = [
|
315 |
"A capybara wearing a suit holding a sign that reads Hello World",
|
|
|
317 |
]
|
318 |
gr.Examples(examples=examples, inputs=[prompt])
|
319 |
|
320 |
+
def update_gallery():
|
321 |
+
"""Update the file gallery"""
|
322 |
+
files = get_generated_images()
|
323 |
+
return [
|
324 |
+
(f["path"], f"{f['name']}\n{f['date']}")
|
325 |
+
for f in files
|
326 |
+
]
|
327 |
|
|
|
328 |
@spaces.GPU(duration=600)
|
329 |
def generate_all(prompt, negative_prompt, seed, randomize_seed, width, height, guidance_scale, num_inference_steps, progress=gr.Progress()):
|
330 |
outputs = []
|
|
|
339 |
print(f"IMAGE GENERATED {model_name} {update_memory_usage()}")
|
340 |
outputs.extend([image, used_seed])
|
341 |
|
|
|
|
|
|
|
342 |
except Exception as e:
|
343 |
outputs.extend([None, None])
|
344 |
print(f"Error generating with {model_name}: {str(e)}")
|
345 |
|
346 |
+
# Update the gallery after generation
|
347 |
+
gallery_images = update_gallery()
|
348 |
+
file_gallery.update(value=gallery_images)
|
349 |
+
|
350 |
return outputs
|
351 |
|
|
|
352 |
output_components = []
|
353 |
for model_name in MODEL_CONFIGS.keys():
|
354 |
output_components.extend([results[model_name], seeds[model_name]])
|
|
|
368 |
outputs=output_components,
|
369 |
)
|
370 |
|
371 |
+
refresh_button.click(
|
372 |
+
fn=update_gallery,
|
373 |
+
inputs=[],
|
374 |
+
outputs=[file_gallery],
|
375 |
+
)
|
376 |
+
|
377 |
+
demo.load(
|
378 |
+
fn=update_gallery,
|
379 |
+
inputs=[],
|
380 |
+
outputs=[file_gallery],
|
381 |
+
)
|
382 |
+
|
383 |
if __name__ == "__main__":
|
384 |
demo.launch()
|