Update app.py
Browse files
app.py
CHANGED
@@ -11,12 +11,27 @@ from diffusers.utils import load_image, export_to_video
|
|
11 |
from PIL import Image
|
12 |
from huggingface_hub import hf_hub_download
|
13 |
|
|
|
|
|
|
|
14 |
|
|
|
|
|
|
|
|
|
|
|
|
|
15 |
pipe = StableVideoDiffusionPipeline.from_pretrained(
|
16 |
-
"stabilityai/stable-video-diffusion-img2vid-xt", torch_dtype=
|
17 |
)
|
18 |
-
pipe.to(
|
19 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
20 |
max_64_bit_int = 2**63 - 1
|
21 |
|
22 |
# Function to sample video from the input image
|
@@ -29,7 +44,6 @@ def sample(
|
|
29 |
version: str = "svd_xt",
|
30 |
cond_aug: float = 0.02,
|
31 |
decoding_t: int = 3, # Number of frames decoded at a time! This eats most VRAM. Reduce if necessary.
|
32 |
-
device: str = "cuda",
|
33 |
output_folder: str = "outputs",
|
34 |
):
|
35 |
if image.mode == "RGBA":
|
@@ -42,20 +56,30 @@ def sample(
|
|
42 |
os.makedirs(output_folder, exist_ok=True)
|
43 |
base_count = len(glob(os.path.join(output_folder, "*.mp4")))
|
44 |
video_path = os.path.join(output_folder, f"{base_count:06d}.mp4")
|
45 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
46 |
export_to_video(frames, video_path, fps=fps_id)
|
47 |
torch.manual_seed(seed)
|
48 |
return video_path, seed
|
49 |
|
50 |
-
# Function to resize the uploaded image
|
51 |
def resize_image(image, output_size=(1024, 576)):
|
|
|
52 |
target_aspect = output_size[0] / output_size[1]
|
53 |
image_aspect = image.width / image.height
|
54 |
|
55 |
if image_aspect > target_aspect:
|
56 |
new_height = output_size[1]
|
57 |
new_width = int(new_height * image_aspect)
|
58 |
-
resized_image = image.resize((new_width, new_height), Image.LANCZOS)
|
59 |
left = (new_width - output_size[0]) / 2
|
60 |
top = 0
|
61 |
right = (new_width + output_size[0]) / 2
|
@@ -63,7 +87,7 @@ def resize_image(image, output_size=(1024, 576)):
|
|
63 |
else:
|
64 |
new_width = output_size[0]
|
65 |
new_height = int(new_width / image_aspect)
|
66 |
-
resized_image = image.resize((new_width, new_height), Image.LANCZOS)
|
67 |
left = 0
|
68 |
top = (new_height - output_size[1]) / 2
|
69 |
right = output_size[0]
|
@@ -75,39 +99,50 @@ def resize_image(image, output_size=(1024, 576)):
|
|
75 |
# Dynamically load image files from the 'images' directory
|
76 |
def get_example_images():
|
77 |
image_dir = "images/"
|
|
|
|
|
78 |
image_files = glob(os.path.join(image_dir, "*.png")) + glob(os.path.join(image_dir, "*.jpg"))
|
79 |
return image_files
|
80 |
|
81 |
# Gradio interface setup
|
82 |
with gr.Blocks() as demo:
|
83 |
-
gr.Markdown('''# Stable Video Diffusion
|
84 |
-
|
85 |
|
86 |
with gr.Row():
|
87 |
with gr.Column():
|
88 |
-
image = gr.Image(label="Upload
|
89 |
-
generate_btn = gr.Button("Generate")
|
90 |
-
video = gr.Video()
|
91 |
|
92 |
-
with gr.Accordion("Advanced
|
93 |
-
seed = gr.Slider(label="Seed", value=42,
|
94 |
-
randomize_seed = gr.Checkbox(label="Randomize
|
95 |
-
motion_bucket_id = gr.Slider(label="Motion
|
96 |
-
fps_id = gr.Slider(label="Frames
|
97 |
|
|
|
98 |
image.upload(fn=resize_image, inputs=image, outputs=image, queue=False)
|
99 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
100 |
|
101 |
# Dynamically load examples from the filesystem
|
102 |
example_images = get_example_images()
|
103 |
-
|
104 |
-
|
105 |
-
|
106 |
-
|
107 |
-
|
108 |
-
|
109 |
-
|
|
|
110 |
|
111 |
if __name__ == "__main__":
|
112 |
demo.queue(max_size=20)
|
113 |
-
demo.launch(share=True)
|
|
|
11 |
from PIL import Image
|
12 |
from huggingface_hub import hf_hub_download
|
13 |
|
14 |
+
# ------------------------------------------------------------------------
|
15 |
+
# FIX: Adapt to the available hardware (GPU or CPU)
|
16 |
+
# ------------------------------------------------------------------------
|
17 |
|
18 |
+
# Automatically detect the device and select the appropriate data type.
|
19 |
+
# This makes the code runnable on machines with or without a dedicated NVIDIA GPU.
|
20 |
+
device = "cuda" if torch.cuda.is_available() else "cpu"
|
21 |
+
torch_dtype = torch.float16 if device == "cuda" else torch.float32
|
22 |
+
|
23 |
+
# Load the pipeline onto the detected device.
|
24 |
pipe = StableVideoDiffusionPipeline.from_pretrained(
|
25 |
+
"stabilityai/stable-video-diffusion-img2vid-xt", torch_dtype=torch_dtype, variant="fp16"
|
26 |
)
|
27 |
+
pipe.to(device)
|
28 |
+
|
29 |
+
# Apply torch.compile for optimization only if on a GPU, as it's most effective there.
|
30 |
+
if device == "cuda":
|
31 |
+
pipe.unet = torch.compile(pipe.unet, mode="reduce-overhead", fullgraph=True)
|
32 |
+
|
33 |
+
# ------------------------------------------------------------------------
|
34 |
+
|
35 |
max_64_bit_int = 2**63 - 1
|
36 |
|
37 |
# Function to sample video from the input image
|
|
|
44 |
version: str = "svd_xt",
|
45 |
cond_aug: float = 0.02,
|
46 |
decoding_t: int = 3, # Number of frames decoded at a time! This eats most VRAM. Reduce if necessary.
|
|
|
47 |
output_folder: str = "outputs",
|
48 |
):
|
49 |
if image.mode == "RGBA":
|
|
|
56 |
os.makedirs(output_folder, exist_ok=True)
|
57 |
base_count = len(glob(os.path.join(output_folder, "*.mp4")))
|
58 |
video_path = os.path.join(output_folder, f"{base_count:06d}.mp4")
|
59 |
+
|
60 |
+
frames = pipe(
|
61 |
+
image,
|
62 |
+
decode_chunk_size=decoding_t,
|
63 |
+
generator=generator,
|
64 |
+
motion_bucket_id=motion_bucket_id,
|
65 |
+
noise_aug_strength=0.1,
|
66 |
+
num_frames=25
|
67 |
+
).frames[0]
|
68 |
+
|
69 |
export_to_video(frames, video_path, fps=fps_id)
|
70 |
torch.manual_seed(seed)
|
71 |
return video_path, seed
|
72 |
|
73 |
+
# Function to resize the uploaded image to the model's optimal input size
|
74 |
def resize_image(image, output_size=(1024, 576)):
|
75 |
+
# Resizes and crops the image to a 16:9 aspect ratio.
|
76 |
target_aspect = output_size[0] / output_size[1]
|
77 |
image_aspect = image.width / image.height
|
78 |
|
79 |
if image_aspect > target_aspect:
|
80 |
new_height = output_size[1]
|
81 |
new_width = int(new_height * image_aspect)
|
82 |
+
resized_image = image.resize((new_width, new_height), Image.Resampling.LANCZOS)
|
83 |
left = (new_width - output_size[0]) / 2
|
84 |
top = 0
|
85 |
right = (new_width + output_size[0]) / 2
|
|
|
87 |
else:
|
88 |
new_width = output_size[0]
|
89 |
new_height = int(new_width / image_aspect)
|
90 |
+
resized_image = image.resize((new_width, new_height), Image.Resampling.LANCZOS)
|
91 |
left = 0
|
92 |
top = (new_height - output_size[1]) / 2
|
93 |
right = output_size[0]
|
|
|
99 |
# Dynamically load image files from the 'images' directory
|
100 |
def get_example_images():
|
101 |
image_dir = "images/"
|
102 |
+
if not os.path.exists(image_dir):
|
103 |
+
os.makedirs(image_dir)
|
104 |
image_files = glob(os.path.join(image_dir, "*.png")) + glob(os.path.join(image_dir, "*.jpg"))
|
105 |
return image_files
|
106 |
|
107 |
# Gradio interface setup
|
108 |
with gr.Blocks() as demo:
|
109 |
+
gr.Markdown('''# Stable Video Diffusion
|
110 |
+
#### Generate short videos from a single image.''')
|
111 |
|
112 |
with gr.Row():
|
113 |
with gr.Column():
|
114 |
+
image = gr.Image(label="Upload Your Image", type="pil")
|
115 |
+
generate_btn = gr.Button("Generate Video", variant="primary")
|
116 |
+
video = gr.Video(label="Generated Video")
|
117 |
|
118 |
+
with gr.Accordion("Advanced Options", open=False):
|
119 |
+
seed = gr.Slider(label="Seed", value=42, minimum=0, maximum=max_64_bit_int, step=1)
|
120 |
+
randomize_seed = gr.Checkbox(label="Randomize Seed", value=True)
|
121 |
+
motion_bucket_id = gr.Slider(label="Motion Bucket ID", info="Controls the amount of motion in the video.", value=127, minimum=1, maximum=255)
|
122 |
+
fps_id = gr.Slider(label="Frames Per Second (FPS)", info="Adjusts the playback speed of the video.", value=7, minimum=5, maximum=30)
|
123 |
|
124 |
+
# When a new image is uploaded, process it immediately
|
125 |
image.upload(fn=resize_image, inputs=image, outputs=image, queue=False)
|
126 |
+
|
127 |
+
# When the generate button is clicked, run the sampling function
|
128 |
+
generate_btn.click(
|
129 |
+
fn=sample,
|
130 |
+
inputs=[image, seed, randomize_seed, motion_bucket_id, fps_id],
|
131 |
+
outputs=[video, seed],
|
132 |
+
api_name="video"
|
133 |
+
)
|
134 |
|
135 |
# Dynamically load examples from the filesystem
|
136 |
example_images = get_example_images()
|
137 |
+
if example_images:
|
138 |
+
gr.Examples(
|
139 |
+
examples=example_images,
|
140 |
+
inputs=image,
|
141 |
+
outputs=[video, seed],
|
142 |
+
fn=lambda img: sample(resize_image(Image.open(img))), # Resize example images before sampling
|
143 |
+
cache_examples=True,
|
144 |
+
)
|
145 |
|
146 |
if __name__ == "__main__":
|
147 |
demo.queue(max_size=20)
|
148 |
+
demo.launch(share=True)
|