Commit
•
36be800
1
Parent(s):
8010ebe
Update app.py
Browse files
app.py
CHANGED
@@ -16,6 +16,8 @@ from huggingface_hub import hf_hub_download
|
|
16 |
|
17 |
#gradio.helpers.CACHED_FOLDER = '/data/cache'
|
18 |
|
|
|
|
|
19 |
pipe = StableVideoDiffusionPipeline.from_pretrained(
|
20 |
"stabilityai/stable-video-diffusion-img2vid-xt", torch_dtype=torch.float16, variant="fp16"
|
21 |
)
|
@@ -36,7 +38,14 @@ def sample(
|
|
36 |
decoding_t: int = 3, # Number of frames decoded at a time! This eats most VRAM. Reduce if necessary.
|
37 |
device: str = "cuda",
|
38 |
output_folder: str = "outputs",
|
|
|
39 |
):
|
|
|
|
|
|
|
|
|
|
|
|
|
40 |
if image.mode == "RGBA":
|
41 |
image = image.convert("RGB")
|
42 |
|
@@ -89,6 +98,11 @@ with gr.Blocks() as demo:
|
|
89 |
gr.Markdown('''# Community demo for Stable Video Diffusion - Img2Vid - XT ([model](https://huggingface.co/stabilityai/stable-video-diffusion-img2vid-xt), [paper](https://stability.ai/research/stable-video-diffusion-scaling-latent-video-diffusion-models-to-large-datasets), [stability's ui waitlist](https://stability.ai/contact))
|
90 |
#### Research release ([_non-commercial_](https://huggingface.co/stabilityai/stable-video-diffusion-img2vid-xt/blob/main/LICENSE)): generate `4s` vid from a single image at (`25 frames` at `6 fps`). this demo uses [🧨 diffusers for low VRAM and fast generation](https://huggingface.co/docs/diffusers/main/en/using-diffusers/svd).
|
91 |
''')
|
|
|
|
|
|
|
|
|
|
|
92 |
with gr.Row():
|
93 |
with gr.Column():
|
94 |
image = gr.Image(label="Upload your image", type="pil")
|
@@ -100,25 +114,8 @@ with gr.Blocks() as demo:
|
|
100 |
motion_bucket_id = gr.Slider(label="Motion bucket id", info="Controls how much motion to add/remove from the image", value=127, minimum=1, maximum=255)
|
101 |
fps_id = gr.Slider(label="Frames per second", info="The length of your video in seconds will be 25/fps", value=6, minimum=5, maximum=30)
|
102 |
|
103 |
-
image.upload(fn=resize_image, inputs=image, outputs=image, queue=False)
|
104 |
-
generate_btn.click(fn=sample, inputs=[image, seed, randomize_seed, motion_bucket_id, fps_id], outputs=[video, seed], api_name="video")
|
105 |
-
gr.Examples(
|
106 |
-
examples=[
|
107 |
-
"images/blink_meme.png",
|
108 |
-
"images/confused2_meme.png",
|
109 |
-
"images/disaster_meme.png",
|
110 |
-
"images/distracted_meme.png",
|
111 |
-
"images/hide_meme.png",
|
112 |
-
"images/nazare_meme.png",
|
113 |
-
"images/success_meme.png",
|
114 |
-
"images/willy_meme.png",
|
115 |
-
"images/wink_meme.png"
|
116 |
-
],
|
117 |
-
inputs=image,
|
118 |
-
outputs=[video, seed],
|
119 |
-
fn=sample,
|
120 |
-
cache_examples=True,
|
121 |
-
)
|
122 |
|
123 |
if __name__ == "__main__":
|
124 |
demo.queue(max_size=20)
|
|
|
16 |
|
17 |
#gradio.helpers.CACHED_FOLDER = '/data/cache'
|
18 |
|
19 |
+
SECRET_TOKEN = os.getenv('SECRET_TOKEN', 'default_secret')
|
20 |
+
|
21 |
pipe = StableVideoDiffusionPipeline.from_pretrained(
|
22 |
"stabilityai/stable-video-diffusion-img2vid-xt", torch_dtype=torch.float16, variant="fp16"
|
23 |
)
|
|
|
38 |
decoding_t: int = 3, # Number of frames decoded at a time! This eats most VRAM. Reduce if necessary.
|
39 |
device: str = "cuda",
|
40 |
output_folder: str = "outputs",
|
41 |
+
secret_token: string = "",
|
42 |
):
|
43 |
+
if secret_token != SECRET_TOKEN:
|
44 |
+
raise gr.Error(
|
45 |
+
f'Invalid secret token. Please fork the original space if you want to use it for yourself.')
|
46 |
+
|
47 |
+
image = resize_image(image)
|
48 |
+
|
49 |
if image.mode == "RGBA":
|
50 |
image = image.convert("RGB")
|
51 |
|
|
|
98 |
gr.Markdown('''# Community demo for Stable Video Diffusion - Img2Vid - XT ([model](https://huggingface.co/stabilityai/stable-video-diffusion-img2vid-xt), [paper](https://stability.ai/research/stable-video-diffusion-scaling-latent-video-diffusion-models-to-large-datasets), [stability's ui waitlist](https://stability.ai/contact))
|
99 |
#### Research release ([_non-commercial_](https://huggingface.co/stabilityai/stable-video-diffusion-img2vid-xt/blob/main/LICENSE)): generate `4s` vid from a single image at (`25 frames` at `6 fps`). this demo uses [🧨 diffusers for low VRAM and fast generation](https://huggingface.co/docs/diffusers/main/en/using-diffusers/svd).
|
100 |
''')
|
101 |
+
secret_token = gr.Text(
|
102 |
+
label='Secret Token',
|
103 |
+
max_lines=1,
|
104 |
+
placeholder='Enter your secret token',
|
105 |
+
)
|
106 |
with gr.Row():
|
107 |
with gr.Column():
|
108 |
image = gr.Image(label="Upload your image", type="pil")
|
|
|
114 |
motion_bucket_id = gr.Slider(label="Motion bucket id", info="Controls how much motion to add/remove from the image", value=127, minimum=1, maximum=255)
|
115 |
fps_id = gr.Slider(label="Frames per second", info="The length of your video in seconds will be 25/fps", value=6, minimum=5, maximum=30)
|
116 |
|
117 |
+
# image.upload(fn=resize_image, inputs=image, outputs=image, queue=False)
|
118 |
+
generate_btn.click(fn=sample, inputs=[image, seed, randomize_seed, motion_bucket_id, fps_id, secret_token], outputs=[video, seed], api_name="video")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
119 |
|
120 |
if __name__ == "__main__":
|
121 |
demo.queue(max_size=20)
|