Spaces:
Running
on
T4
Running
on
T4
Update app.py
Browse files
app.py
CHANGED
|
@@ -49,27 +49,12 @@ with gr.Blocks() as demo:
|
|
| 49 |
<h2 style="text-align: center;">DreamTalk: When Expressive Talking Head Generation Meets Diffusion Probabilistic Models</h2>
|
| 50 |
<p style="text-align: center;">
|
| 51 |
DreamTalk is a diffusion-based audio-driven expressive talking head generation framework that can produce high-quality talking head videos across diverse speaking styles. DreamTalk exhibits robust performance with a diverse array of inputs, including songs, speech in multiple languages, noisy audio, and out-of-domain portraits.
|
| 52 |
-
<img src="https://github.com/ali-vilab/dreamtalk/raw/main/media/teaser.gif" />
|
| 53 |
</p>
|
| 54 |
""")
|
| 55 |
with gr.Row():
|
| 56 |
with gr.Column():
|
| 57 |
image_path = gr.Image(label="Image", type="filepath", sources=["upload"])
|
| 58 |
-
gr.Examples(
|
| 59 |
-
examples = [
|
| 60 |
-
"data/src_img/uncropped/face3.png",
|
| 61 |
-
"data/src_img/uncropped/male_face.png",
|
| 62 |
-
"data/src_img/uncropped/uncut_src_img.jpg",
|
| 63 |
-
"data/src_img/cropped/chpa5.png",
|
| 64 |
-
"data/src_img/cropped/cut_img.png",
|
| 65 |
-
"data/src_img/cropped/f30.png",
|
| 66 |
-
"data/src_img/cropped/menglu2.png",
|
| 67 |
-
"data/src_img/cropped/nscu2.png",
|
| 68 |
-
"data/src_img/cropped/zp1.png",
|
| 69 |
-
"data/src_img/cropped/zt12.png"
|
| 70 |
-
],
|
| 71 |
-
inputs=[image_path]
|
| 72 |
-
)
|
| 73 |
audio_input = gr.Audio(label="Audio input", type="filepath", sources=["upload"], value="data/audio/acknowledgement_english.m4a")
|
| 74 |
with gr.Row():
|
| 75 |
audio_list = gr.Dropdown(
|
|
@@ -122,6 +107,21 @@ with gr.Blocks() as demo:
|
|
| 122 |
],
|
| 123 |
value = "M030_front_neutral_level1_001.mat"
|
| 124 |
)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 125 |
run_btn = gr.Button("Run")
|
| 126 |
with gr.Column():
|
| 127 |
output_video = gr.Video(format="mp4")
|
|
|
|
| 49 |
<h2 style="text-align: center;">DreamTalk: When Expressive Talking Head Generation Meets Diffusion Probabilistic Models</h2>
|
| 50 |
<p style="text-align: center;">
|
| 51 |
DreamTalk is a diffusion-based audio-driven expressive talking head generation framework that can produce high-quality talking head videos across diverse speaking styles. DreamTalk exhibits robust performance with a diverse array of inputs, including songs, speech in multiple languages, noisy audio, and out-of-domain portraits.
|
| 52 |
+
<img src="https://github.com/ali-vilab/dreamtalk/raw/main/media/teaser.gif" style="margin: 0 auto" width="40%" />
|
| 53 |
</p>
|
| 54 |
""")
|
| 55 |
with gr.Row():
|
| 56 |
with gr.Column():
|
| 57 |
image_path = gr.Image(label="Image", type="filepath", sources=["upload"])
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 58 |
audio_input = gr.Audio(label="Audio input", type="filepath", sources=["upload"], value="data/audio/acknowledgement_english.m4a")
|
| 59 |
with gr.Row():
|
| 60 |
audio_list = gr.Dropdown(
|
|
|
|
| 107 |
],
|
| 108 |
value = "M030_front_neutral_level1_001.mat"
|
| 109 |
)
|
| 110 |
+
gr.Examples(
|
| 111 |
+
examples = [
|
| 112 |
+
"data/src_img/uncropped/face3.png",
|
| 113 |
+
"data/src_img/uncropped/male_face.png",
|
| 114 |
+
"data/src_img/uncropped/uncut_src_img.jpg",
|
| 115 |
+
"data/src_img/cropped/chpa5.png",
|
| 116 |
+
"data/src_img/cropped/cut_img.png",
|
| 117 |
+
"data/src_img/cropped/f30.png",
|
| 118 |
+
"data/src_img/cropped/menglu2.png",
|
| 119 |
+
"data/src_img/cropped/nscu2.png",
|
| 120 |
+
"data/src_img/cropped/zp1.png",
|
| 121 |
+
"data/src_img/cropped/zt12.png"
|
| 122 |
+
],
|
| 123 |
+
inputs=[image_path]
|
| 124 |
+
)
|
| 125 |
run_btn = gr.Button("Run")
|
| 126 |
with gr.Column():
|
| 127 |
output_video = gr.Video(format="mp4")
|