File size: 9,504 Bytes
44189a1 dc78df8 ea0a1d3 dc78df8 ea0a1d3 dc78df8 0424d1d dc78df8 0424d1d dc78df8 25fc9d6 dc78df8 25fc9d6 dc78df8 0424d1d dc78df8 0424d1d dc78df8 2e26875 dc78df8 64d39dd dc78df8 25fc9d6 dc78df8 0424d1d dc78df8 0424d1d dc78df8 2e26875 dc78df8 25fc9d6 dc78df8 44189a1 dc78df8 25fc9d6 dc78df8 d3480fc 44189a1 dc78df8 44189a1 dc78df8 64d39dd dc78df8 44189a1 dc78df8 |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 |
from gradio_imageslider import ImageSlider
import functools
import os
import tempfile
import diffusers
import gradio as gr
import imageio as imageio
import numpy as np
import spaces
import torch as torch
from PIL import Image
from tqdm import tqdm
from pathlib import Path
import gradio
from gradio.utils import get_cache_folder
from infer import lotus, lotus_video
import transformers
transformers.utils.move_cache()
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
def infer(path_input, seed):
name_base, name_ext = os.path.splitext(os.path.basename(path_input))
output_g, output_d = lotus(path_input, 'depth', seed, device)
if not os.path.exists("files/output"):
os.makedirs("files/output")
g_save_path = os.path.join("files/output", f"{name_base}_g{name_ext}")
d_save_path = os.path.join("files/output", f"{name_base}_d{name_ext}")
output_g.save(g_save_path)
output_d.save(d_save_path)
return [path_input, g_save_path], [path_input, d_save_path]
def infer_video(path_input, seed):
frames_g, frames_d = lotus_video(path_input, 'depth', seed, device)
if not os.path.exists("files/output"):
os.makedirs("files/output")
name_base, _ = os.path.splitext(os.path.basename(path_input))
g_save_path = os.path.join("files/output", f"{name_base}_g.mp4")
d_save_path = os.path.join("files/output", f"{name_base}_d.mp4")
imageio.mimsave(g_save_path, frames_g)
imageio.mimsave(d_save_path, frames_d)
return [g_save_path, d_save_path]
def run_demo_server():
infer_gpu = spaces.GPU(functools.partial(infer, seed=0))
infer_video_gpu = spaces.GPU(functools.partial(infer_video, seed=0))
gradio_theme = gr.themes.Default()
with gr.Blocks(
theme=gradio_theme,
title="LOTUS (Depth)",
css="""
#download {
height: 118px;
}
.slider .inner {
width: 5px;
background: #FFF;
}
.viewport {
aspect-ratio: 4/3;
}
.tabs button.selected {
font-size: 20px !important;
color: crimson !important;
}
h1 {
text-align: center;
display: block;
}
h2 {
text-align: center;
display: block;
}
h3 {
text-align: center;
display: block;
}
.md_feedback li {
margin-bottom: 0px !important;
}
""",
head="""
<script async src="https://www.googletagmanager.com/gtag/js?id=G-1FWSVCGZTG"></script>
<script>
window.dataLayer = window.dataLayer || [];
function gtag() {dataLayer.push(arguments);}
gtag('js', new Date());
gtag('config', 'G-1FWSVCGZTG');
</script>
""",
) as demo:
gr.Markdown(
"""
# LOTUS: Diffusion-based Visual Foundation Model for High-quality Dense Prediction
<p align="center">
<a title="Page" href="https://lotus3d.github.io/" target="_blank" rel="noopener noreferrer" style="display: inline-block;">
<img src="https://img.shields.io/badge/Project-Website-pink?logo=googlechrome&logoColor=white">
</a>
<a title="arXiv" href="https://arxiv.org/abs/2409.18124" target="_blank" rel="noopener noreferrer" style="display: inline-block;">
<img src="https://img.shields.io/badge/arXiv-Paper-b31b1b?logo=arxiv&logoColor=white">
</a>
<a title="Github" href="https://github.com/EnVision-Research/Lotus" target="_blank" rel="noopener noreferrer" style="display: inline-block;">
<img src="https://img.shields.io/github/stars/EnVision-Research/Lotus?label=GitHub%20%E2%98%85&logo=github&color=C8C" alt="badge-github-stars">
</a>
<a title="Social" href="https://x.com/haodongli00/status/1839524569058582884" target="_blank" rel="noopener noreferrer" style="display: inline-block;">
<img src="https://www.obukhov.ai/img/badges/badge-social.svg" alt="social">
</a>
"""
)
with gr.Tabs(elem_classes=["tabs"]):
with gr.Tab("IMAGE"):
with gr.Row():
with gr.Column():
image_input = gr.Image(
label="Input Image",
type="filepath",
)
seed = gr.Number(
label="Seed (only for Generative mode)",
minimum=0,
maximum=999999999,
)
with gr.Row():
image_submit_btn = gr.Button(
value="Predict Depth!", variant="primary"
)
image_reset_btn = gr.Button(value="Reset")
with gr.Column():
image_output_g = ImageSlider(
label="Output (Generative)",
type="filepath",
interactive=False,
elem_classes="slider",
position=0.25,
)
with gr.Row():
image_output_d = ImageSlider(
label="Output (Discriminative)",
type="filepath",
interactive=False,
elem_classes="slider",
position=0.25,
)
gr.Examples(
fn=infer_gpu,
examples=sorted([
[os.path.join("files", "images", name), 0]
for name in os.listdir(os.path.join("files", "images"))
]),
inputs=[image_input, seed],
outputs=[image_output_g, image_output_d],
cache_examples=False,
)
with gr.Tab("VIDEO"):
with gr.Row():
with gr.Column():
input_video = gr.Video(
label="Input Video",
autoplay=True,
loop=True,
)
seed = gr.Number(
label="Seed (only for Generative mode)",
minimum=0,
maximum=999999999,
)
with gr.Row():
video_submit_btn = gr.Button(
value="Predict Depth!", variant="primary"
)
video_reset_btn = gr.Button(value="Reset")
with gr.Column():
video_output_g = gr.Video(
label="Output (Generative)",
interactive=False,
autoplay=True,
loop=True,
show_share_button=True,
)
with gr.Row():
video_output_d = gr.Video(
label="Output (Discriminative)",
interactive=False,
autoplay=True,
loop=True,
show_share_button=True,
)
gr.Examples(
fn=infer_video_gpu,
examples=sorted([
[os.path.join("files", "videos", name), 0]
for name in os.listdir(os.path.join("files", "videos"))
]),
inputs=[input_video, seed],
outputs=[video_output_g, video_output_d],
cache_examples=False,
)
### Image
image_submit_btn.click(
fn=infer_gpu,
inputs=[image_input, seed],
outputs=[image_output_g, image_output_d],
concurrency_limit=1,
)
image_reset_btn.click(
fn=lambda: (
None,
None,
None,
),
inputs=[],
outputs=[image_output_g, image_output_d],
queue=False,
)
### Video
video_submit_btn.click(
fn=infer_video_gpu,
inputs=[input_video, seed],
outputs=[video_output_g, video_output_d],
queue=True,
)
video_reset_btn.click(
fn=lambda: (None, None, None),
inputs=[],
outputs=[video_output_g, video_output_d],
)
### Server launch
demo.queue(
api_open=False,
).launch(
server_name="0.0.0.0",
server_port=7860,
)
def main():
os.system("pip freeze")
if os.path.exists("files/output"):
os.system("rm -rf files/output")
run_demo_server()
if __name__ == "__main__":
main()
|