# Copyright 2024 Anton Obukhov, ETH Zurich. All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# --------------------------------------------------------------------------
# If you find this code useful, we kindly ask you to cite our paper in your work.
# Please find bibtex at: https://github.com/prs-eth/Marigold#-citation
# More information about the method can be found at https://marigoldmonodepth.github.io
# --------------------------------------------------------------------------
import functools
import os
import sys
import tempfile
import av
import numpy as np
import spaces
import gradio as gr
import torch as torch
import einops
from huggingface_hub import login
from colorize import colorize_depth_multi_thread
from video_io import get_video_fps, write_video_from_numpy
VERBOSE = False
MAX_FRAMES = 100
def process(pipe, device, path_input):
print(f"Processing {path_input}")
path_output_dir = tempfile.mkdtemp()
os.makedirs(path_output_dir, exist_ok=True)
name_base = os.path.splitext(os.path.basename(path_input))[0]
path_out_in = os.path.join(path_output_dir, f"{name_base}_depth_input.mp4")
path_out_vis = os.path.join(path_output_dir, f"{name_base}_depth_colored.mp4")
output_fps = int(get_video_fps(path_input))
container = av.open(path_input)
stream = container.streams.video[0]
fps = float(stream.average_rate)
duration_sec = float(stream.duration * stream.time_base) if stream.duration else 0
total_frames = int(duration_sec * fps)
if total_frames > MAX_FRAMES:
gr.Warning(
f"Only the first {MAX_FRAMES} frames (~{MAX_FRAMES / fps:.1f} sec.) will be processed for demonstration; "
f"use the code from GitHub for full processing"
)
generator = torch.Generator(device=device)
generator.manual_seed(2024)
pipe_out: RollingDepthOutput = pipe(
# input setting
input_video_path=path_input,
start_frame=0,
frame_count=min(MAX_FRAMES, total_frames), # 0 = all
processing_res=768,
# infer setting
dilations=[1, 25],
cap_dilation=True,
snippet_lengths=[3],
init_infer_steps=[1],
strides=[1],
coalign_kwargs=None,
refine_step=0, # 0 = off
max_vae_bs=8, # batch size for encoder/decoder
# other settings
generator=generator,
verbose=VERBOSE,
# output settings
restore_res=False,
unload_snippet=False,
)
depth_pred = pipe_out.depth_pred # [N 1 H W]
# Colorize results
cmap = "Spectral_r"
colored_np = colorize_depth_multi_thread(
depth=depth_pred.numpy(),
valid_mask=None,
chunk_size=4,
num_threads=4,
color_map=cmap,
verbose=VERBOSE,
) # [n h w 3], in [0, 255]
write_video_from_numpy(
frames=colored_np,
output_path=path_out_vis,
fps=output_fps,
crf=23,
preset="medium",
verbose=VERBOSE,
)
# Save rgb
rgb = (pipe_out.input_rgb.numpy() * 255).astype(np.uint8) # [N 3 H W]
rgb = einops.rearrange(rgb, "n c h w -> n h w c")
write_video_from_numpy(
frames=rgb,
output_path=path_out_in,
fps=output_fps,
crf=23,
preset="medium",
verbose=VERBOSE,
)
return path_out_in, path_out_vis
def run_demo_server(pipe, device):
process_pipe = spaces.GPU(functools.partial(process, pipe, device), duration=120)
os.environ["GRADIO_ALLOW_FLAGGING"] = "never"
with gr.Blocks(
analytics_enabled=False,
title="RollingDepth",
css="""
h1 {
text-align: center;
display: block;
}
h2 {
text-align: center;
display: block;
}
h3 {
text-align: center;
display: block;
}
""",
) as demo:
gr.HTML(
"""
🛹 RollingDepth 🛹: Video Depth without Video Models
RollingDepth is the state-of-the-art depth estimator for videos in the wild. Upload your video into the
left pane, or click any of the examples below. The result preview will be computed and
appear in the right panes. For full functionality, use the code on GitHub.
TIP: When running out of GPU time, fork the demo.
"""
)
with gr.Row(equal_height=True):
with gr.Column(scale=1):
input_video = gr.Video(label="Input Video")
with gr.Column(scale=2):
with gr.Row(equal_height=True):
output_video_1 = gr.Video(
label="Preprocessed video",
interactive=False,
autoplay=True,
loop=True,
show_share_button=True,
scale=5,
)
output_video_2 = gr.Video(
label="Generated Depth Video",
interactive=False,
autoplay=True,
loop=True,
show_share_button=True,
scale=5,
)
with gr.Row(equal_height=True):
with gr.Column(scale=1):
with gr.Row(equal_height=False):
generate_btn = gr.Button("Generate")
with gr.Column(scale=2):
pass
gr.Examples(
examples=[
["files/gokart.mp4"],
["files/horse.mp4"],
["files/walking.mp4"],
],
inputs=[input_video],
outputs=[output_video_1, output_video_2],
fn=process_pipe,
cache_examples=True,
cache_mode="eager",
)
generate_btn.click(
fn=process_pipe,
inputs=[input_video],
outputs=[output_video_1, output_video_2],
)
demo.queue(
api_open=False,
).launch(
server_name="0.0.0.0",
server_port=7860,
)
def main():
os.system("pip freeze")
os.system("pip uninstall -y diffusers")
os.system("pip install rollingdepth_src/diffusers")
os.system("pip freeze")
if "HF_TOKEN_LOGIN" in os.environ:
login(token=os.environ["HF_TOKEN_LOGIN"])
if torch.cuda.is_available():
device = torch.device("cuda")
elif torch.backends.mps.is_available():
device = torch.device("mps")
else:
device = torch.device("cpu")
sys.path.append(os.path.join(os.path.dirname(__file__), "rollingdepth_src"))
from rollingdepth import RollingDepthOutput, RollingDepthPipeline
pipe: RollingDepthPipeline = RollingDepthPipeline.from_pretrained(
"prs-eth/rollingdepth-v1-0",
torch_dtype=torch.float16,
)
pipe.set_progress_bar_config(disable=True)
try:
import xformers
pipe.enable_xformers_memory_efficient_attention()
except:
pass # run without xformers
pipe = pipe.to(device)
run_demo_server(pipe, device)
if __name__ == "__main__":
main()
