Spaces:
Runtime error
Runtime error
隐藏huggingface上的gradio UI,当后端服务启用
Browse files
app.py
CHANGED
@@ -68,44 +68,44 @@ def remove_tips():
|
|
68 |
return gr.update(visible=False)
|
69 |
|
70 |
|
71 |
-
def get_example():
|
72 |
-
|
73 |
-
|
74 |
-
|
75 |
-
|
76 |
-
|
77 |
-
|
78 |
-
|
79 |
-
|
80 |
-
|
81 |
-
|
82 |
-
|
83 |
-
|
84 |
-
|
85 |
-
|
86 |
-
|
87 |
-
|
88 |
-
|
89 |
-
|
90 |
-
|
91 |
-
|
92 |
-
|
93 |
-
|
94 |
-
|
95 |
-
|
96 |
-
|
97 |
-
|
98 |
-
|
99 |
-
|
100 |
-
|
101 |
-
|
102 |
-
|
103 |
-
|
104 |
-
|
105 |
-
|
106 |
-
|
107 |
-
def run_for_examples(face_file, prompt, style, negative_prompt):
|
108 |
-
|
109 |
|
110 |
|
111 |
def convert_from_cv2_to_image(img: np.ndarray) -> Image:
|
@@ -240,6 +240,7 @@ def generate_image(
|
|
240 |
|
241 |
width, height = face_kps.size
|
242 |
|
|
|
243 |
if enhance_face_region:
|
244 |
control_mask = np.zeros([height, width, 3])
|
245 |
x1, y1, x2, y2 = face_info["bbox"]
|
@@ -251,8 +252,8 @@ def generate_image(
|
|
251 |
|
252 |
generator = torch.Generator(device=device).manual_seed(seed)
|
253 |
|
254 |
-
print("Start inference...")
|
255 |
-
print(f"[Debug] Prompt: {prompt}, \n[Debug] Neg Prompt: {negative_prompt}")
|
256 |
|
257 |
pipe.set_ip_adapter_scale(adapter_strength_ratio)
|
258 |
images = pipe(
|
@@ -272,169 +273,169 @@ def generate_image(
|
|
272 |
return images[0], gr.update(visible=True)
|
273 |
|
274 |
|
275 |
-
### Description
|
276 |
-
title = r"""
|
277 |
-
<h1 align="center">InstantID: Zero-shot Identity-Preserving Generation in Seconds</h1>
|
278 |
-
"""
|
279 |
-
|
280 |
-
description = r"""
|
281 |
-
<b>Official 🤗 Gradio demo</b> for <a href='https://github.com/InstantID/InstantID' target='_blank'><b>InstantID: Zero-shot Identity-Preserving Generation in Seconds</b></a>.<br>
|
282 |
-
|
283 |
-
How to use:<br>
|
284 |
-
1. Upload a person image. For multiple person images, we will only detect the biggest face. Make sure face is not too small and not significantly blocked or blurred.
|
285 |
-
2. (Optionally) upload another person image as reference pose. If not uploaded, we will use the first person image to extract landmarks. If you use a cropped face at step1, it is recommeneded to upload it to extract a new pose.
|
286 |
-
3. Enter a text prompt as done in normal text-to-image models.
|
287 |
-
4. Click the <b>Submit</b> button to start customizing.
|
288 |
-
5. Share your customizd photo with your friends, enjoy😊!
|
289 |
-
"""
|
290 |
-
|
291 |
-
article = r"""
|
292 |
-
---
|
293 |
-
📝 **Citation**
|
294 |
-
<br>
|
295 |
-
If our work is helpful for your research or applications, please cite us via:
|
296 |
-
```bibtex
|
297 |
-
@article{wang2024instantid,
|
298 |
-
|
299 |
-
|
300 |
-
|
301 |
-
|
302 |
-
}
|
303 |
-
```
|
304 |
-
📧 **Contact**
|
305 |
-
<br>
|
306 |
-
If you have any questions, please feel free to open an issue or directly reach us out at <b>[email protected]</b>.
|
307 |
-
"""
|
308 |
-
|
309 |
-
tips = r"""
|
310 |
-
### Usage tips of InstantID
|
311 |
-
1. If you're unsatisfied with the similarity, increase the weight of controlnet_conditioning_scale (IdentityNet) and ip_adapter_scale (Adapter).
|
312 |
-
2. If the generated image is over-saturated, decrease the ip_adapter_scale. If not work, decrease controlnet_conditioning_scale.
|
313 |
-
3. If text control is not as expected, decrease ip_adapter_scale.
|
314 |
-
4. Find a good base model always makes a difference.
|
315 |
-
"""
|
316 |
-
|
317 |
-
css = """
|
318 |
-
.gradio-container {width: 85% !important}
|
319 |
-
"""
|
320 |
-
with gr.Blocks(css=css) as demo:
|
321 |
-
|
322 |
-
|
323 |
-
|
324 |
-
|
325 |
-
|
326 |
-
|
327 |
-
|
328 |
-
|
329 |
-
|
330 |
-
|
331 |
-
|
332 |
-
|
333 |
-
|
334 |
-
|
335 |
-
|
336 |
-
|
337 |
-
|
338 |
-
|
339 |
-
|
340 |
-
|
341 |
-
|
342 |
-
|
343 |
-
|
344 |
-
|
345 |
-
|
346 |
-
|
347 |
-
|
348 |
-
|
349 |
-
|
350 |
-
|
351 |
-
|
352 |
-
|
353 |
-
|
354 |
-
|
355 |
-
|
356 |
-
|
357 |
-
|
358 |
-
|
359 |
-
|
360 |
-
|
361 |
-
|
362 |
-
|
363 |
-
|
364 |
-
|
365 |
-
|
366 |
-
|
367 |
-
|
368 |
-
|
369 |
-
|
370 |
-
|
371 |
-
|
372 |
-
|
373 |
-
|
374 |
-
|
375 |
-
|
376 |
-
|
377 |
-
|
378 |
-
|
379 |
-
|
380 |
-
|
381 |
-
|
382 |
-
|
383 |
-
|
384 |
-
|
385 |
-
|
386 |
-
|
387 |
-
|
388 |
-
|
389 |
-
|
390 |
-
|
391 |
-
|
392 |
-
|
393 |
-
|
394 |
-
|
395 |
-
|
396 |
-
|
397 |
-
|
398 |
-
|
399 |
-
|
400 |
-
|
401 |
-
|
402 |
-
|
403 |
-
|
404 |
-
|
405 |
-
|
406 |
-
|
407 |
-
|
408 |
-
|
409 |
-
|
410 |
-
|
411 |
-
|
412 |
-
|
413 |
-
|
414 |
-
|
415 |
-
|
416 |
-
|
417 |
-
|
418 |
-
|
419 |
-
|
420 |
-
|
421 |
-
|
422 |
-
|
423 |
-
|
424 |
-
|
425 |
-
|
426 |
-
|
427 |
-
|
428 |
-
|
429 |
-
gr.Examples(
|
430 |
-
|
431 |
-
|
432 |
-
|
433 |
-
|
434 |
-
|
435 |
-
)
|
436 |
-
|
437 |
-
gr.Markdown(article)
|
438 |
|
439 |
demo.queue(api_open=False)
|
440 |
demo.launch()
|
|
|
68 |
return gr.update(visible=False)
|
69 |
|
70 |
|
71 |
+
# def get_example():
|
72 |
+
# case = [
|
73 |
+
# [
|
74 |
+
# "./examples/yann-lecun_resize.jpg",
|
75 |
+
# "a man",
|
76 |
+
# "Snow",
|
77 |
+
# "(lowres, low quality, worst quality:1.2), (text:1.2), watermark, (frame:1.2), deformed, ugly, deformed eyes, blur, out of focus, blurry, deformed cat, deformed, photo, anthropomorphic cat, monochrome, photo, pet collar, gun, weapon, blue, 3d, drones, drone, buildings in background, green",
|
78 |
+
# ],
|
79 |
+
# [
|
80 |
+
# "./examples/musk_resize.jpeg",
|
81 |
+
# "a man",
|
82 |
+
# "Mars",
|
83 |
+
# "(lowres, low quality, worst quality:1.2), (text:1.2), watermark, (frame:1.2), deformed, ugly, deformed eyes, blur, out of focus, blurry, deformed cat, deformed, photo, anthropomorphic cat, monochrome, photo, pet collar, gun, weapon, blue, 3d, drones, drone, buildings in background, green",
|
84 |
+
# ],
|
85 |
+
# [
|
86 |
+
# "./examples/sam_resize.png",
|
87 |
+
# "a man",
|
88 |
+
# "Jungle",
|
89 |
+
# "(lowres, low quality, worst quality:1.2), (text:1.2), watermark, (frame:1.2), deformed, ugly, deformed eyes, blur, out of focus, blurry, deformed cat, deformed, photo, anthropomorphic cat, monochrome, photo, pet collar, gun, weapon, blue, 3d, drones, drone, buildings in background, gree",
|
90 |
+
# ],
|
91 |
+
# [
|
92 |
+
# "./examples/schmidhuber_resize.png",
|
93 |
+
# "a man",
|
94 |
+
# "Neon",
|
95 |
+
# "(lowres, low quality, worst quality:1.2), (text:1.2), watermark, (frame:1.2), deformed, ugly, deformed eyes, blur, out of focus, blurry, deformed cat, deformed, photo, anthropomorphic cat, monochrome, photo, pet collar, gun, weapon, blue, 3d, drones, drone, buildings in background, green",
|
96 |
+
# ],
|
97 |
+
# [
|
98 |
+
# "./examples/kaifu_resize.png",
|
99 |
+
# "a man",
|
100 |
+
# "Vibrant Color",
|
101 |
+
# "(lowres, low quality, worst quality:1.2), (text:1.2), watermark, (frame:1.2), deformed, ugly, deformed eyes, blur, out of focus, blurry, deformed cat, deformed, photo, anthropomorphic cat, monochrome, photo, pet collar, gun, weapon, blue, 3d, drones, drone, buildings in background, green",
|
102 |
+
# ],
|
103 |
+
# ]
|
104 |
+
# return case
|
105 |
+
|
106 |
+
|
107 |
+
# def run_for_examples(face_file, prompt, style, negative_prompt):
|
108 |
+
# return generate_image(face_file, None, prompt, negative_prompt, style, True, 30, 0.8, 0.8, 5, 42)
|
109 |
|
110 |
|
111 |
def convert_from_cv2_to_image(img: np.ndarray) -> Image:
|
|
|
240 |
|
241 |
width, height = face_kps.size
|
242 |
|
243 |
+
# 面部增强
|
244 |
if enhance_face_region:
|
245 |
control_mask = np.zeros([height, width, 3])
|
246 |
x1, y1, x2, y2 = face_info["bbox"]
|
|
|
252 |
|
253 |
generator = torch.Generator(device=device).manual_seed(seed)
|
254 |
|
255 |
+
# print("Start inference...")
|
256 |
+
# print(f"[Debug] Prompt: {prompt}, \n[Debug] Neg Prompt: {negative_prompt}")
|
257 |
|
258 |
pipe.set_ip_adapter_scale(adapter_strength_ratio)
|
259 |
images = pipe(
|
|
|
273 |
return images[0], gr.update(visible=True)
|
274 |
|
275 |
|
276 |
+
# ### Description
|
277 |
+
# title = r"""
|
278 |
+
# <h1 align="center">InstantID: Zero-shot Identity-Preserving Generation in Seconds</h1>
|
279 |
+
# """
|
280 |
+
|
281 |
+
# description = r"""
|
282 |
+
# <b>Official 🤗 Gradio demo</b> for <a href='https://github.com/InstantID/InstantID' target='_blank'><b>InstantID: Zero-shot Identity-Preserving Generation in Seconds</b></a>.<br>
|
283 |
+
|
284 |
+
# How to use:<br>
|
285 |
+
# 1. Upload a person image. For multiple person images, we will only detect the biggest face. Make sure face is not too small and not significantly blocked or blurred.
|
286 |
+
# 2. (Optionally) upload another person image as reference pose. If not uploaded, we will use the first person image to extract landmarks. If you use a cropped face at step1, it is recommeneded to upload it to extract a new pose.
|
287 |
+
# 3. Enter a text prompt as done in normal text-to-image models.
|
288 |
+
# 4. Click the <b>Submit</b> button to start customizing.
|
289 |
+
# 5. Share your customizd photo with your friends, enjoy😊!
|
290 |
+
# """
|
291 |
+
|
292 |
+
# article = r"""
|
293 |
+
# ---
|
294 |
+
# 📝 **Citation**
|
295 |
+
# <br>
|
296 |
+
# If our work is helpful for your research or applications, please cite us via:
|
297 |
+
# ```bibtex
|
298 |
+
# @article{wang2024instantid,
|
299 |
+
# title={InstantID: Zero-shot Identity-Preserving Generation in Seconds},
|
300 |
+
# author={Wang, Qixun and Bai, Xu and Wang, Haofan and Qin, Zekui and Chen, Anthony},
|
301 |
+
# journal={arXiv preprint arXiv:2401.07519},
|
302 |
+
# year={2024}
|
303 |
+
# }
|
304 |
+
# ```
|
305 |
+
# 📧 **Contact**
|
306 |
+
# <br>
|
307 |
+
# If you have any questions, please feel free to open an issue or directly reach us out at <b>[email protected]</b>.
|
308 |
+
# """
|
309 |
+
|
310 |
+
# tips = r"""
|
311 |
+
# ### Usage tips of InstantID
|
312 |
+
# 1. If you're unsatisfied with the similarity, increase the weight of controlnet_conditioning_scale (IdentityNet) and ip_adapter_scale (Adapter).
|
313 |
+
# 2. If the generated image is over-saturated, decrease the ip_adapter_scale. If not work, decrease controlnet_conditioning_scale.
|
314 |
+
# 3. If text control is not as expected, decrease ip_adapter_scale.
|
315 |
+
# 4. Find a good base model always makes a difference.
|
316 |
+
# """
|
317 |
+
|
318 |
+
# css = """
|
319 |
+
# .gradio-container {width: 85% !important}
|
320 |
+
# """
|
321 |
+
# with gr.Blocks(css=css) as demo:
|
322 |
+
# # description
|
323 |
+
# gr.Markdown(title)
|
324 |
+
# gr.Markdown(description)
|
325 |
+
|
326 |
+
# with gr.Row():
|
327 |
+
# with gr.Column():
|
328 |
+
# # upload face image
|
329 |
+
# face_file = gr.Image(label="Upload a photo of your face", type="filepath")
|
330 |
+
|
331 |
+
# # optional: upload a reference pose image
|
332 |
+
# pose_file = gr.Image(label="Upload a reference pose image (optional)", type="filepath")
|
333 |
+
|
334 |
+
# # prompt
|
335 |
+
# prompt = gr.Textbox(
|
336 |
+
# label="Prompt",
|
337 |
+
# info="Give simple prompt is enough to achieve good face fedility",
|
338 |
+
# placeholder="A photo of a person",
|
339 |
+
# value="",
|
340 |
+
# )
|
341 |
+
|
342 |
+
# submit = gr.Button("Submit", variant="primary")
|
343 |
+
|
344 |
+
# style = gr.Dropdown(label="Style template", choices=STYLE_NAMES, value=DEFAULT_STYLE_NAME)
|
345 |
+
|
346 |
+
# # strength
|
347 |
+
# identitynet_strength_ratio = gr.Slider(
|
348 |
+
# label="IdentityNet strength (for fedility)",
|
349 |
+
# minimum=0,
|
350 |
+
# maximum=1.5,
|
351 |
+
# step=0.05,
|
352 |
+
# value=0.80,
|
353 |
+
# )
|
354 |
+
# adapter_strength_ratio = gr.Slider(
|
355 |
+
# label="Image adapter strength (for detail)",
|
356 |
+
# minimum=0,
|
357 |
+
# maximum=1.5,
|
358 |
+
# step=0.05,
|
359 |
+
# value=0.80,
|
360 |
+
# )
|
361 |
+
|
362 |
+
# with gr.Accordion(open=False, label="Advanced Options"):
|
363 |
+
# negative_prompt = gr.Textbox(
|
364 |
+
# label="Negative Prompt",
|
365 |
+
# placeholder="low quality",
|
366 |
+
# value="(lowres, low quality, worst quality:1.2), (text:1.2), watermark, (frame:1.2), deformed, ugly, deformed eyes, blur, out of focus, blurry, deformed cat, deformed, nudity,naked, bikini, skimpy, scanty, bare skin, lingerie, swimsuit, exposed, see-through, photo, anthropomorphic cat, monochrome, pet collar, gun, weapon, blue, 3d, drones, drone, buildings in background, green",
|
367 |
+
# )
|
368 |
+
# num_steps = gr.Slider(
|
369 |
+
# label="Number of sample steps",
|
370 |
+
# minimum=20,
|
371 |
+
# maximum=100,
|
372 |
+
# step=1,
|
373 |
+
# value=30,
|
374 |
+
# )
|
375 |
+
# guidance_scale = gr.Slider(
|
376 |
+
# label="Guidance scale",
|
377 |
+
# minimum=0.1,
|
378 |
+
# maximum=10.0,
|
379 |
+
# step=0.1,
|
380 |
+
# value=5,
|
381 |
+
# )
|
382 |
+
# seed = gr.Slider(
|
383 |
+
# label="Seed",
|
384 |
+
# minimum=0,
|
385 |
+
# maximum=MAX_SEED,
|
386 |
+
# step=1,
|
387 |
+
# value=42,
|
388 |
+
# )
|
389 |
+
# randomize_seed = gr.Checkbox(label="Randomize seed", value=True)
|
390 |
+
# enhance_face_region = gr.Checkbox(label="Enhance non-face region", value=True)
|
391 |
+
|
392 |
+
# with gr.Column():
|
393 |
+
# output_image = gr.Image(label="Generated Image")
|
394 |
+
# usage_tips = gr.Markdown(label="Usage tips of InstantID", value=tips, visible=False)
|
395 |
+
|
396 |
+
# submit.click(
|
397 |
+
# fn=remove_tips,
|
398 |
+
# outputs=usage_tips,
|
399 |
+
# queue=False,
|
400 |
+
# api_name=False,
|
401 |
+
# ).then(
|
402 |
+
# fn=randomize_seed_fn,
|
403 |
+
# inputs=[seed, randomize_seed],
|
404 |
+
# outputs=seed,
|
405 |
+
# queue=False,
|
406 |
+
# api_name=False,
|
407 |
+
# ).then(
|
408 |
+
# fn=check_input_image,
|
409 |
+
# inputs=face_file,
|
410 |
+
# queue=False,
|
411 |
+
# api_name=False,
|
412 |
+
# ).success(
|
413 |
+
# fn=generate_image,
|
414 |
+
# inputs=[
|
415 |
+
# face_file,
|
416 |
+
# pose_file,
|
417 |
+
# prompt,
|
418 |
+
# negative_prompt,
|
419 |
+
# style,
|
420 |
+
# enhance_face_region,
|
421 |
+
# num_steps,
|
422 |
+
# identitynet_strength_ratio,
|
423 |
+
# adapter_strength_ratio,
|
424 |
+
# guidance_scale,
|
425 |
+
# seed,
|
426 |
+
# ],
|
427 |
+
# outputs=[output_image, usage_tips],
|
428 |
+
# )
|
429 |
+
|
430 |
+
# gr.Examples(
|
431 |
+
# examples=get_example(),
|
432 |
+
# inputs=[face_file, prompt, style, negative_prompt],
|
433 |
+
# outputs=[output_image, usage_tips],
|
434 |
+
# fn=run_for_examples,
|
435 |
+
# cache_examples=True,
|
436 |
+
# )
|
437 |
+
|
438 |
+
# gr.Markdown(article)
|
439 |
|
440 |
demo.queue(api_open=False)
|
441 |
demo.launch()
|