-
Notifications
You must be signed in to change notification settings - Fork 37
/
webui_emotions.py
215 lines (171 loc) · 8.04 KB
/
webui_emotions.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
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
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
import gradio as gr
import os
from code_for_webui.download_models_openxlab import download
# Download models and check for exists
# download()
from code_for_webui.demo_EDTalk_A_using_predefined_exp_weights import Demo as Demo_EDTalk_A_using_predefined_exp_weights
from code_for_webui.demo_lip_pose import Demo as Demo_lip_pose
demo_lip_pose = Demo_lip_pose()
demo_EDTalk_A_using_predefined_exp_weights = Demo_EDTalk_A_using_predefined_exp_weights()
def get_example():
case = [
[
'res/results_by_facesr/demo_EDTalk_A.png',
False,
"res/results_by_facesr/demo_EDTalk_A.wav",
"test_data/pose_source1.mp4",
False,
"I don't wanna generate emotional expression",
True,
],
[
'test_data/identity_source.jpg',
False,
"test_data/mouth_source.wav",
"res/results_by_facesr/demo_EDTalk_A.mp4",
False,
"surprised",
True,
],
[
'test_data/uncrop_face.jpg',
True,
"test_data/test/11.wav",
"test_data/uncrop_Obama.mp4",
True,
"happy",
True,
]
]
return case
tips = r"""
### Usage tips of EDTalk
1. If you're not satisfied with the results, check whether the image or the video is cropped.
2. If you're not satisfied with the results, check whether "Use Face Super-Resolution" is selected.
3. If you're not satisfied with the results, choose a pose video with similar head pose to source image.
"""
def run_inference(source_image, need_crop_source_img, audio_file, pose_video, need_crop_pose_video, exp_type, face_sr):
# Get file paths
try:
source_path = source_image if source_image else ""
audio_driving_path = audio_file if audio_file else ""
pose_driving_path = pose_video if pose_video else "test_data/pose_source1.mp4"
# Construct the command
if exp_type in ["angry", "contempt", "disgusted", "fear", "happy", "sad", "surprised"]:
# command = (
# f"python demo_EDTalk_A_using_predefined_exp_weights.py --source_path {source_path} "
# f"--audio_driving_path {audio_driving_path} --pose_driving_path {pose_driving_path} "
# f"--exp_type {exp_type} --save_path {save_path}"
# )
demo_EDTalk_A_using_predefined_exp_weights.process_data(source_path, pose_driving_path, audio_driving_path, exp_type, need_crop_source_img, need_crop_pose_video, face_sr)
save_path = demo_EDTalk_A_using_predefined_exp_weights.run()
else:
demo_lip_pose.process_data(source_path, pose_driving_path, audio_driving_path, need_crop_source_img, need_crop_pose_video, face_sr)
save_path = demo_lip_pose.run()
save_512_path = save_path.replace('.mp4','_512.mp4')
# Check if the output video file exists
if not os.path.exists(save_path):
return None, gr.Markdown("Error: Video generation failed. Please check your inputs and try again.")
if face_sr == False:
print("here face_sr == False:")
return gr.Video(value=save_path), None, gr.Markdown("Video (256*256 only) generated successfully!")
elif os.path.exists(save_512_path):
print("here os.path.exists(save_512_path):")
return gr.Video(value=save_path), gr.Video(value=save_512_path), gr.Markdown(tips)
else:
print("else")
return None, None, gr.Markdown("Video generated failed, please retry it.")
except Exception as e:
print(str(e))
return None, None, gr.Markdown("Video generated failed, please retry it.")
# return f"Output saved to: {save_path}"
def remove_tips():
return gr.update(visible=False)
def main():
### Description
title = r"""
<h1 align="center">EDTalk: Efficient Disentanglement for Emotional Talking Head Synthesis</h1>
"""
description = r"""
<b>Official 🤗 Gradio demo</b> for <a href='https://github.com/tanshuai0219/EDTalk' target='_blank'><b>EDTalk: Efficient Disentanglement for Emotional Talking Head Synthesis</b></a>.<br>
How to use:<br>
1. Upload an image with a face. Ensure the face is not too small and is clearly visible without significant obstructions or blurring.
2. If the face is not be cropped by crop_image2.py, please click "Crop the Source Image"
3. Upload a video for head pose source. Ensure the face is not too small and is clearly visible without significant obstructions or blurring.
4. If the face is not be cropped by crop_video.py, please click "Crop the Pose Video"
5. Upload an audio
6. Choose the exp type
7. (Recommended) click the "Use Face Super-Resolution"
8. Share the generated videos with your friends and enjoy! 😊
"""
article = r"""
---
📝 **Citation**
<br>
If our work is helpful for your research or applications, please cite us via:
```bibtex
@inproceedings{tan2024edtalk,
title = {EDTalk: Efficient Disentanglement for Emotional Talking Head Synthesis},
author = {Tan, Shuai and Ji, Bin and Bi, Mengxiao and Pan, Ye},
booktitle = {Proceedings of the European Conference on Computer Vision (ECCV)},
year = {2024}
}
```
📧 **Contact**
<br>
If you have any questions, please feel free to open an issue or directly reach us out at <b>[email protected]</b>.
"""
tips = r"""
### Usage tips of EDTalk
1. If you're not satisfied with the results, check whether the image or the video is cropped.
2. If you're not satisfied with the results, check whether "Use Face Super-Resolution" is selected.
3. If you're not satisfied with the results, choose a pose video with similar head pose to source image.
"""
css = '''
.gradio-container {width: 85% !important}
'''
with gr.Blocks(css=css) as demo:
# description
gr.Markdown(title)
gr.Markdown(description)
with gr.Row():
with gr.Column():
# upload face image
source_file = gr.Image(type="filepath",label="Select Source Image.")
crop_image = gr.Checkbox(label="Crop the Source Image")
driving_audio = gr.Audio(type="filepath", label="Select Audio File")
pose_video = gr.Video(label="Select Pose Video. If no pose video, generate video will have the default pose")
crop_video = gr.Checkbox(label="Crop the Pose Video")
exp_type = gr.Dropdown(
choices=["I don't wanna generate emotional expression","angry", "contempt", "disgusted", "fear", "happy", "sad", "surprised"],
label="Select Expression Type",
value="I don't wanna generate emotional expression"
)
face_sr = gr.Checkbox(label="Use Face Super-Resolution")
submit = gr.Button("Submit", variant="primary")
with gr.Column():
output_256 = gr.Video(label="Generated Video (256)")
output_512 = gr.Video(label="Generated Video (512)")
output_log = gr.Markdown(label="Usage tips of EDTalk", value=tips ,visible=False)
submit.click(
fn=remove_tips,
outputs=output_log,
).then(
fn=run_inference,
inputs=[source_file, crop_image, driving_audio, pose_video, crop_video, exp_type, face_sr],
outputs=[output_256, output_512, output_log]
)
gr.Examples(
examples=get_example(),
inputs=[source_file, crop_image, driving_audio, pose_video, crop_video, exp_type, face_sr],
run_on_click=True,
fn=run_inference,
outputs=[output_256, output_512, output_log],
cache_examples=True,
)
gr.Markdown(article)
demo.launch()
# # Launch the interface
# iface.launch()
if __name__ == "__main__":
main()