Spaces:
Running
on
Zero
Running
on
Zero
Update app.py
Browse files
app.py
CHANGED
|
@@ -202,11 +202,14 @@ def run_lora(prompt, cfg_scale, steps, selected_index, randomize_seed, seed, asp
|
|
| 202 |
# Load LoRAs based on speed mode
|
| 203 |
if speed_mode == "Speed (4 steps)":
|
| 204 |
with calculateDuration("Loading Lightning LoRA and style LoRA"):
|
|
|
|
| 205 |
pipe.load_lora_weights(
|
| 206 |
LIGHTNING_LORA_REPO,
|
| 207 |
weight_name=LIGHTNING_LORA_WEIGHT,
|
| 208 |
adapter_name="lightning"
|
| 209 |
)
|
|
|
|
|
|
|
| 210 |
weight_name = selected_lora.get("weights", None)
|
| 211 |
pipe.load_lora_weights(
|
| 212 |
lora_path,
|
|
@@ -214,14 +217,19 @@ def run_lora(prompt, cfg_scale, steps, selected_index, randomize_seed, seed, asp
|
|
| 214 |
low_cpu_mem_usage=True,
|
| 215 |
adapter_name="style"
|
| 216 |
)
|
|
|
|
|
|
|
| 217 |
pipe.set_adapters(["lightning", "style"], adapter_weights=[1.0, lora_scale])
|
| 218 |
elif speed_mode == "Speed (8 steps)":
|
| 219 |
with calculateDuration("Loading Lightning LoRA and style LoRA"):
|
|
|
|
| 220 |
pipe.load_lora_weights(
|
| 221 |
LIGHTNING_LORA_REPO,
|
| 222 |
weight_name=LIGHTNING8_LORA_WEIGHT,
|
| 223 |
adapter_name="lightning"
|
| 224 |
)
|
|
|
|
|
|
|
| 225 |
weight_name = selected_lora.get("weights", None)
|
| 226 |
pipe.load_lora_weights(
|
| 227 |
lora_path,
|
|
@@ -229,8 +237,11 @@ def run_lora(prompt, cfg_scale, steps, selected_index, randomize_seed, seed, asp
|
|
| 229 |
low_cpu_mem_usage=True,
|
| 230 |
adapter_name="style"
|
| 231 |
)
|
|
|
|
|
|
|
| 232 |
pipe.set_adapters(["lightning", "style"], adapter_weights=[1.0, lora_scale])
|
| 233 |
else:
|
|
|
|
| 234 |
with calculateDuration(f"Loading LoRA weights for {selected_lora['title']}"):
|
| 235 |
weight_name = selected_lora.get("weights", None)
|
| 236 |
pipe.load_lora_weights(
|
|
@@ -254,4 +265,289 @@ def run_lora(prompt, cfg_scale, steps, selected_index, randomize_seed, seed, asp
|
|
| 254 |
|
| 255 |
return final_image, seed
|
| 256 |
|
| 257 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 202 |
# Load LoRAs based on speed mode
|
| 203 |
if speed_mode == "Speed (4 steps)":
|
| 204 |
with calculateDuration("Loading Lightning LoRA and style LoRA"):
|
| 205 |
+
# Load Lightning LoRA first
|
| 206 |
pipe.load_lora_weights(
|
| 207 |
LIGHTNING_LORA_REPO,
|
| 208 |
weight_name=LIGHTNING_LORA_WEIGHT,
|
| 209 |
adapter_name="lightning"
|
| 210 |
)
|
| 211 |
+
|
| 212 |
+
# Load the selected style LoRA
|
| 213 |
weight_name = selected_lora.get("weights", None)
|
| 214 |
pipe.load_lora_weights(
|
| 215 |
lora_path,
|
|
|
|
| 217 |
low_cpu_mem_usage=True,
|
| 218 |
adapter_name="style"
|
| 219 |
)
|
| 220 |
+
|
| 221 |
+
# Set both adapters active with their weights
|
| 222 |
pipe.set_adapters(["lightning", "style"], adapter_weights=[1.0, lora_scale])
|
| 223 |
elif speed_mode == "Speed (8 steps)":
|
| 224 |
with calculateDuration("Loading Lightning LoRA and style LoRA"):
|
| 225 |
+
# Load Lightning LoRA first
|
| 226 |
pipe.load_lora_weights(
|
| 227 |
LIGHTNING_LORA_REPO,
|
| 228 |
weight_name=LIGHTNING8_LORA_WEIGHT,
|
| 229 |
adapter_name="lightning"
|
| 230 |
)
|
| 231 |
+
|
| 232 |
+
# Load the selected style LoRA
|
| 233 |
weight_name = selected_lora.get("weights", None)
|
| 234 |
pipe.load_lora_weights(
|
| 235 |
lora_path,
|
|
|
|
| 237 |
low_cpu_mem_usage=True,
|
| 238 |
adapter_name="style"
|
| 239 |
)
|
| 240 |
+
|
| 241 |
+
# Set both adapters active with their weights
|
| 242 |
pipe.set_adapters(["lightning", "style"], adapter_weights=[1.0, lora_scale])
|
| 243 |
else:
|
| 244 |
+
# Quality mode - only load the style LoRA
|
| 245 |
with calculateDuration(f"Loading LoRA weights for {selected_lora['title']}"):
|
| 246 |
weight_name = selected_lora.get("weights", None)
|
| 247 |
pipe.load_lora_weights(
|
|
|
|
| 265 |
|
| 266 |
return final_image, seed
|
| 267 |
|
| 268 |
+
def get_huggingface_safetensors(link):
|
| 269 |
+
split_link = link.split("/")
|
| 270 |
+
if len(split_link) != 2:
|
| 271 |
+
raise Exception("Invalid Hugging Face repository link format.")
|
| 272 |
+
|
| 273 |
+
print(f"Repository attempted: {split_link}")
|
| 274 |
+
|
| 275 |
+
# Load model card
|
| 276 |
+
model_card = ModelCard.load(link)
|
| 277 |
+
base_model = model_card.data.get("base_model")
|
| 278 |
+
print(f"Base model: {base_model}")
|
| 279 |
+
|
| 280 |
+
# Validate model type (for Qwen-Image)
|
| 281 |
+
acceptable_models = {"Qwen/Qwen-Image"}
|
| 282 |
+
|
| 283 |
+
models_to_check = base_model if isinstance(base_model, list) else [base_model]
|
| 284 |
+
|
| 285 |
+
if not any(model in acceptable_models for model in models_to_check):
|
| 286 |
+
raise Exception("Not a Qwen-Image LoRA!")
|
| 287 |
+
|
| 288 |
+
# Extract image and trigger word
|
| 289 |
+
image_path = model_card.data.get("widget", [{}])[0].get("output", {}).get("url", None)
|
| 290 |
+
trigger_word = model_card.data.get("instance_prompt", "")
|
| 291 |
+
image_url = f"https://huggingface.co/{link}/resolve/main/{image_path}" if image_path else None
|
| 292 |
+
|
| 293 |
+
# Initialize Hugging Face file system
|
| 294 |
+
fs = HfFileSystem()
|
| 295 |
+
try:
|
| 296 |
+
list_of_files = fs.ls(link, detail=False)
|
| 297 |
+
|
| 298 |
+
# Find safetensors file
|
| 299 |
+
safetensors_name = None
|
| 300 |
+
for file in list_of_files:
|
| 301 |
+
filename = file.split("/")[-1]
|
| 302 |
+
if filename.endswith(".safetensors"):
|
| 303 |
+
safetensors_name = filename
|
| 304 |
+
break
|
| 305 |
+
|
| 306 |
+
if not safetensors_name:
|
| 307 |
+
raise Exception("No valid *.safetensors file found in the repository.")
|
| 308 |
+
|
| 309 |
+
except Exception as e:
|
| 310 |
+
print(e)
|
| 311 |
+
raise Exception("You didn't include a valid Hugging Face repository with a *.safetensors LoRA")
|
| 312 |
+
|
| 313 |
+
return split_link[1], link, safetensors_name, trigger_word, image_url
|
| 314 |
+
|
| 315 |
+
def check_custom_model(link):
|
| 316 |
+
print(f"Checking a custom model on: {link}")
|
| 317 |
+
|
| 318 |
+
if link.endswith('.safetensors'):
|
| 319 |
+
if 'huggingface.co' in link:
|
| 320 |
+
parts = link.split('/')
|
| 321 |
+
try:
|
| 322 |
+
hf_index = parts.index('huggingface.co')
|
| 323 |
+
username = parts[hf_index + 1]
|
| 324 |
+
repo_name = parts[hf_index + 2]
|
| 325 |
+
repo = f"{username}/{repo_name}"
|
| 326 |
+
|
| 327 |
+
safetensors_name = parts[-1]
|
| 328 |
+
|
| 329 |
+
try:
|
| 330 |
+
model_card = ModelCard.load(repo)
|
| 331 |
+
trigger_word = model_card.data.get("instance_prompt", "")
|
| 332 |
+
image_path = model_card.data.get("widget", [{}])[0].get("output", {}).get("url", None)
|
| 333 |
+
image_url = f"https://huggingface.co/{repo}/resolve/main/{image_path}" if image_path else None
|
| 334 |
+
except:
|
| 335 |
+
trigger_word = ""
|
| 336 |
+
image_url = None
|
| 337 |
+
|
| 338 |
+
return repo_name, repo, safetensors_name, trigger_word, image_url
|
| 339 |
+
except:
|
| 340 |
+
raise Exception("Invalid safetensors URL format")
|
| 341 |
+
|
| 342 |
+
if link.startswith("https://"):
|
| 343 |
+
if link.startswith("https://huggingface.co") or link.startswith("https://www.huggingface.co"):
|
| 344 |
+
link_split = link.split("huggingface.co/")
|
| 345 |
+
return get_huggingface_safetensors(link_split[1])
|
| 346 |
+
else:
|
| 347 |
+
return get_huggingface_safetensors(link)
|
| 348 |
+
|
| 349 |
+
def add_custom_lora(custom_lora):
|
| 350 |
+
global loras
|
| 351 |
+
if custom_lora:
|
| 352 |
+
try:
|
| 353 |
+
title, repo, path, trigger_word, image = check_custom_model(custom_lora)
|
| 354 |
+
print(f"Loaded custom LoRA: {repo}")
|
| 355 |
+
|
| 356 |
+
# Get model card examples for custom LoRA
|
| 357 |
+
model_card_examples = ""
|
| 358 |
+
try:
|
| 359 |
+
model_card = ModelCard.load(repo)
|
| 360 |
+
widget_data = model_card.data.get("widget", [])
|
| 361 |
+
if widget_data and len(widget_data) > 0:
|
| 362 |
+
examples_html = '<div style="margin-top: 10px;">'
|
| 363 |
+
examples_html += '<h4 style="margin-bottom: 8px; font-size: 0.9em;">Sample Images:</h4>'
|
| 364 |
+
examples_html += '<div style="display: grid; grid-template-columns: repeat(4, 1fr); gap: 8px;">'
|
| 365 |
+
|
| 366 |
+
for i, example in enumerate(widget_data[:4]):
|
| 367 |
+
if "output" in example and "url" in example["output"]:
|
| 368 |
+
image_url = f"https://huggingface.co/{repo}/resolve/main/{example['output']['url']}"
|
| 369 |
+
caption = example.get("text", f"Example {i+1}")
|
| 370 |
+
examples_html += f'''
|
| 371 |
+
<div style="text-align: center;">
|
| 372 |
+
<img src="{image_url}" style="width: 100%; height: auto; border-radius: 4px;" />
|
| 373 |
+
<p style="font-size: 0.7em; margin: 2px 0;">{caption[:30]}{'...' if len(caption) > 30 else ''}</p>
|
| 374 |
+
</div>
|
| 375 |
+
'''
|
| 376 |
+
|
| 377 |
+
examples_html += '</div></div>'
|
| 378 |
+
model_card_examples = examples_html
|
| 379 |
+
except Exception as e:
|
| 380 |
+
print(f"Could not load model card examples for custom LoRA: {e}")
|
| 381 |
+
|
| 382 |
+
card = f'''
|
| 383 |
+
<div class="custom_lora_card">
|
| 384 |
+
<span>Loaded custom LoRA:</span>
|
| 385 |
+
<div class="card_internal">
|
| 386 |
+
<img src="{image}" />
|
| 387 |
+
<div>
|
| 388 |
+
<h3>{title}</h3>
|
| 389 |
+
<small>{"Using: <code><b>"+trigger_word+"</code></b> as the trigger word" if trigger_word else "No trigger word found. If there's a trigger word, include it in your prompt"}<br></small>
|
| 390 |
+
</div>
|
| 391 |
+
</div>
|
| 392 |
+
{model_card_examples}
|
| 393 |
+
</div>
|
| 394 |
+
'''
|
| 395 |
+
existing_item_index = next((index for (index, item) in enumerate(loras) if item['repo'] == repo), None)
|
| 396 |
+
if existing_item_index is None:
|
| 397 |
+
new_item = {
|
| 398 |
+
"image": image,
|
| 399 |
+
"title": title,
|
| 400 |
+
"repo": repo,
|
| 401 |
+
"weights": path,
|
| 402 |
+
"trigger_word": trigger_word
|
| 403 |
+
}
|
| 404 |
+
print(new_item)
|
| 405 |
+
loras.append(new_item)
|
| 406 |
+
existing_item_index = len(loras) - 1 # Get the actual index after adding
|
| 407 |
+
|
| 408 |
+
return gr.update(visible=True, value=card), gr.update(visible=True), gr.Gallery(selected_index=None), f"Custom: {path}", existing_item_index, trigger_word
|
| 409 |
+
except Exception as e:
|
| 410 |
+
full_traceback = traceback.format_exc()
|
| 411 |
+
print(f"Full traceback:\n{full_traceback}")
|
| 412 |
+
gr.Warning(f"Invalid LoRA: either you entered an invalid link, or a non-Qwen-Image LoRA, this was the issue: {e}")
|
| 413 |
+
return gr.update(visible=True, value=f"Invalid LoRA: either you entered an invalid link, a non-Qwen-Image LoRA"), gr.update(visible=True), gr.update(), "", None, ""
|
| 414 |
+
else:
|
| 415 |
+
return gr.update(visible=False), gr.update(visible=False), gr.update(), "", None, ""
|
| 416 |
+
|
| 417 |
+
def remove_custom_lora():
|
| 418 |
+
return gr.update(visible=False), gr.update(visible=False), gr.update(), "", None, ""
|
| 419 |
+
|
| 420 |
+
run_lora.zerogpu = True
|
| 421 |
+
|
| 422 |
+
css = '''
|
| 423 |
+
#gen_btn{height: 100%}
|
| 424 |
+
#gen_column{align-self: stretch}
|
| 425 |
+
#title{text-align: center}
|
| 426 |
+
#title h1{font-size: 3em; display:inline-flex; align-items:center}
|
| 427 |
+
#title img{width: 100px; margin-right: 0.5em}
|
| 428 |
+
#gallery .grid-wrap{height: 10vh}
|
| 429 |
+
#lora_list{background: var(--block-background-fill);padding: 0 1em .3em; font-size: 90%}
|
| 430 |
+
.card_internal{display: flex;height: 100px;margin-top: .5em}
|
| 431 |
+
.card_internal img{margin-right: 1em}
|
| 432 |
+
.styler{--form-gap-width: 0px !important}
|
| 433 |
+
#speed_status{padding: .5em; border-radius: 5px; margin: 1em 0}
|
| 434 |
+
'''
|
| 435 |
+
|
| 436 |
+
with gr.Blocks(theme=gr.themes.Soft(), css=css, delete_cache=(60, 60)) as app:
|
| 437 |
+
title = gr.HTML(
|
| 438 |
+
"""<img src=\"https://qianwen-res.oss-cn-beijing.aliyuncs.com/Qwen-Image/qwen_image_logo.png\" alt=\"Qwen-Image\" style=\"width: 280px; margin: 0 auto\">
|
| 439 |
+
<h3 style=\"margin-top: -10px\">LoRA🦜 ChoquinLabs Explorer</h3>""",
|
| 440 |
+
elem_id="title",
|
| 441 |
+
)
|
| 442 |
+
|
| 443 |
+
selected_index = gr.State(None)
|
| 444 |
+
|
| 445 |
+
with gr.Row():
|
| 446 |
+
with gr.Column(scale=3):
|
| 447 |
+
prompt = gr.Textbox(label="Prompt", lines=1, placeholder="Type a prompt after selecting a LoRA")
|
| 448 |
+
with gr.Column(scale=1, elem_id="gen_column"):
|
| 449 |
+
generate_button = gr.Button("Generate", variant="primary", elem_id="gen_btn")
|
| 450 |
+
|
| 451 |
+
with gr.Row():
|
| 452 |
+
with gr.Column():
|
| 453 |
+
selected_info = gr.Markdown("")
|
| 454 |
+
examples_component = gr.Examples(examples=[], inputs=[prompt], label="Sample Prompts", visible=False)
|
| 455 |
+
gallery = gr.Gallery(
|
| 456 |
+
[(item["image"], item["title"]) for item in loras],
|
| 457 |
+
label="LoRA Gallery",
|
| 458 |
+
allow_preview=False,
|
| 459 |
+
columns=3,
|
| 460 |
+
elem_id="gallery",
|
| 461 |
+
show_share_button=False
|
| 462 |
+
)
|
| 463 |
+
with gr.Group():
|
| 464 |
+
custom_lora = gr.Textbox(label="Custom LoRA", info="LoRA Hugging Face path", placeholder="username/qwen-image-custom-lora")
|
| 465 |
+
gr.Markdown("[Check Qwen-Image LoRAs](https://huggingface.co/models?other=base_model:adapter:Qwen/Qwen-Image)", elem_id="lora_list")
|
| 466 |
+
custom_lora_info = gr.HTML(visible=False)
|
| 467 |
+
custom_lora_button = gr.Button("Remove custom LoRA", visible=False)
|
| 468 |
+
|
| 469 |
+
with gr.Column():
|
| 470 |
+
result = gr.Image(label="Generated Image")
|
| 471 |
+
|
| 472 |
+
with gr.Row():
|
| 473 |
+
speed_mode = gr.Radio(
|
| 474 |
+
label="Generation Mode",
|
| 475 |
+
choices=["Speed (4 steps)", "Speed (8 steps)", "Quality (45 steps)"],
|
| 476 |
+
value="Speed (4 steps)",
|
| 477 |
+
info="Speed mode uses Lightning LoRA for faster generation"
|
| 478 |
+
)
|
| 479 |
+
|
| 480 |
+
speed_status = gr.Markdown("Quality mode active", elem_id="speed_status")
|
| 481 |
+
|
| 482 |
+
with gr.Row():
|
| 483 |
+
with gr.Accordion("Advanced Settings", open=False):
|
| 484 |
+
with gr.Column():
|
| 485 |
+
with gr.Row():
|
| 486 |
+
aspect_ratio = gr.Radio(
|
| 487 |
+
label="Aspect Ratio",
|
| 488 |
+
choices=["1:1", "16:9", "9:16", "4:3", "3:4", "3:2", "2:3", "3:1", "2:1"],
|
| 489 |
+
value="16:9"
|
| 490 |
+
)
|
| 491 |
+
|
| 492 |
+
with gr.Row():
|
| 493 |
+
cfg_scale = gr.Slider(
|
| 494 |
+
label="Guidance Scale (True CFG)",
|
| 495 |
+
minimum=1.0,
|
| 496 |
+
maximum=5.0,
|
| 497 |
+
step=0.1,
|
| 498 |
+
value=3.5,
|
| 499 |
+
info="Lower for speed mode, higher for quality"
|
| 500 |
+
)
|
| 501 |
+
steps = gr.Slider(
|
| 502 |
+
label="Steps",
|
| 503 |
+
minimum=4,
|
| 504 |
+
maximum=50,
|
| 505 |
+
step=1,
|
| 506 |
+
value=45,
|
| 507 |
+
info="Automatically set by speed mode"
|
| 508 |
+
)
|
| 509 |
+
|
| 510 |
+
with gr.Row():
|
| 511 |
+
randomize_seed = gr.Checkbox(True, label="Randomize seed")
|
| 512 |
+
seed = gr.Slider(label="Seed", minimum=0, maximum=MAX_SEED, step=1, value=0, randomize=True)
|
| 513 |
+
lora_scale = gr.Slider(label="LoRA Scale", minimum=0, maximum=3, step=0.01, value=1.0)
|
| 514 |
+
|
| 515 |
+
# Event handlers
|
| 516 |
+
gallery.select(
|
| 517 |
+
update_selection,
|
| 518 |
+
inputs=[aspect_ratio],
|
| 519 |
+
outputs=[prompt, selected_info, selected_index, aspect_ratio]
|
| 520 |
+
)
|
| 521 |
+
|
| 522 |
+
speed_mode.change(
|
| 523 |
+
handle_speed_mode,
|
| 524 |
+
inputs=[speed_mode],
|
| 525 |
+
outputs=[speed_status, steps, cfg_scale]
|
| 526 |
+
)
|
| 527 |
+
|
| 528 |
+
custom_lora.input(
|
| 529 |
+
add_custom_lora,
|
| 530 |
+
inputs=[custom_lora],
|
| 531 |
+
outputs=[custom_lora_info, custom_lora_button, gallery, selected_info, selected_index, prompt]
|
| 532 |
+
)
|
| 533 |
+
|
| 534 |
+
custom_lora_button.click(
|
| 535 |
+
remove_custom_lora,
|
| 536 |
+
outputs=[custom_lora_info, custom_lora_button, gallery, selected_info, selected_index, custom_lora]
|
| 537 |
+
)
|
| 538 |
+
|
| 539 |
+
gr.on(
|
| 540 |
+
triggers=[generate_button.click, prompt.submit],
|
| 541 |
+
fn=run_lora,
|
| 542 |
+
inputs=[prompt, cfg_scale, steps, selected_index, randomize_seed, seed, aspect_ratio, lora_scale, speed_mode],
|
| 543 |
+
outputs=[result, seed]
|
| 544 |
+
)
|
| 545 |
+
|
| 546 |
+
app.load(
|
| 547 |
+
fn=handle_speed_mode,
|
| 548 |
+
inputs=[gr.State("Speed (4 steps)")],
|
| 549 |
+
outputs=[speed_status, steps, cfg_scale]
|
| 550 |
+
)
|
| 551 |
+
|
| 552 |
+
app.queue()
|
| 553 |
+
app.launch()
|