Spaces:
Sleeping
Sleeping
Upload folder using huggingface_hub
Browse filesThis view is limited to 50 files because it contains too many changes.
See raw diff
- .gitattributes +17 -0
- .gitignore +137 -0
- README.md +1 -7
- UI.py +81 -0
- app.py +215 -0
- checkpoints/colornet.pth +3 -0
- checkpoints/embed_net.pth +3 -0
- checkpoints/nonlocal_net.pth +3 -0
- cmd.txt +21 -0
- cmd_ddp.txt +20 -0
- docs/.gitignore +0 -0
- environment.yml +0 -0
- examples.zip +3 -0
- examples/bear/ref.jpg +0 -0
- examples/bear/video.mp4 +3 -0
- examples/boat/ref.jpg +0 -0
- examples/boat/video.mp4 +0 -0
- examples/cows/ref.jpg +0 -0
- examples/cows/video.mp4 +3 -0
- examples/flamingo/ref.jpg +0 -0
- examples/flamingo/video.mp4 +3 -0
- gradio_cached_examples/13/log.csv +5 -0
- gradio_cached_examples/13/output/003c3114319372a78bf2f812ebaf0041afa280fb/output_video.mp4 +3 -0
- gradio_cached_examples/13/output/74c76e483235b7e80665e32d7fcdcc3da2be7644/output_video.mp4 +0 -0
- gradio_cached_examples/13/output/7969adca8ae38cb3b38ff8e7bb54688d942c7bc8/output_video.mp4 +3 -0
- gradio_cached_examples/13/output/e6d6153dedeb9fec586b3241311cc49dbc17bc85/output_video.mp4 +0 -0
- inputs/video.mp4/000000000.jpg +0 -0
- inputs/video.mp4/000000001.jpg +0 -0
- inputs/video.mp4/000000002.jpg +0 -0
- inputs/video.mp4/000000003.jpg +0 -0
- inputs/video.mp4/000000004.jpg +0 -0
- inputs/video.mp4/000000005.jpg +0 -0
- inputs/video.mp4/000000006.jpg +0 -0
- inputs/video.mp4/000000007.jpg +0 -0
- inputs/video.mp4/000000008.jpg +0 -0
- inputs/video.mp4/000000009.jpg +0 -0
- inputs/video.mp4/000000010.jpg +0 -0
- inputs/video.mp4/000000011.jpg +0 -0
- inputs/video.mp4/000000012.jpg +0 -0
- inputs/video.mp4/000000013.jpg +0 -0
- inputs/video.mp4/000000014.jpg +0 -0
- inputs/video.mp4/000000015.jpg +0 -0
- inputs/video.mp4/000000016.jpg +0 -0
- inputs/video.mp4/000000017.jpg +0 -0
- inputs/video.mp4/000000018.jpg +0 -0
- inputs/video.mp4/000000019.jpg +0 -0
- inputs/video.mp4/000000020.jpg +0 -0
- inputs/video.mp4/000000021.jpg +0 -0
- inputs/video.mp4/000000022.jpg +0 -0
- inputs/video.mp4/000000023.jpg +0 -0
.gitattributes
CHANGED
@@ -33,3 +33,20 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
|
|
33 |
*.zip filter=lfs diff=lfs merge=lfs -text
|
34 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
35 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
33 |
*.zip filter=lfs diff=lfs merge=lfs -text
|
34 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
35 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
36 |
+
EvalDataset/clips/bear/output_video.mp4 filter=lfs diff=lfs merge=lfs -text
|
37 |
+
EvalDataset/clips/bear/output_video_gray.mp4 filter=lfs diff=lfs merge=lfs -text
|
38 |
+
EvalDataset/clips/boat/output_video_gray.mp4 filter=lfs diff=lfs merge=lfs -text
|
39 |
+
EvalDataset/clips/cows/output_video.mp4 filter=lfs diff=lfs merge=lfs -text
|
40 |
+
EvalDataset/clips/cows/output_video_gray.mp4 filter=lfs diff=lfs merge=lfs -text
|
41 |
+
EvalDataset/clips/dog/output_video.mp4 filter=lfs diff=lfs merge=lfs -text
|
42 |
+
EvalDataset/clips/flamingo/output_video_gray.mp4 filter=lfs diff=lfs merge=lfs -text
|
43 |
+
EvalDataset/ref/goat/0000.jpg filter=lfs diff=lfs merge=lfs -text
|
44 |
+
EvalDataset/ref/hockey/0000.jpg filter=lfs diff=lfs merge=lfs -text
|
45 |
+
EvalDataset/ref/horsejump-high/0000.jpg filter=lfs diff=lfs merge=lfs -text
|
46 |
+
EvalDataset/ref/motorbike/0000.jpg filter=lfs diff=lfs merge=lfs -text
|
47 |
+
EvalDataset/ref/surf/0000.jpg filter=lfs diff=lfs merge=lfs -text
|
48 |
+
examples/bear/video.mp4 filter=lfs diff=lfs merge=lfs -text
|
49 |
+
examples/cows/video.mp4 filter=lfs diff=lfs merge=lfs -text
|
50 |
+
examples/flamingo/video.mp4 filter=lfs diff=lfs merge=lfs -text
|
51 |
+
gradio_cached_examples/13/output/003c3114319372a78bf2f812ebaf0041afa280fb/output_video.mp4 filter=lfs diff=lfs merge=lfs -text
|
52 |
+
gradio_cached_examples/13/output/7969adca8ae38cb3b38ff8e7bb54688d942c7bc8/output_video.mp4 filter=lfs diff=lfs merge=lfs -text
|
.gitignore
ADDED
@@ -0,0 +1,137 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
checkpoints/
|
2 |
+
wandb/
|
3 |
+
.vscode
|
4 |
+
.DS_Store
|
5 |
+
*ckpt*/
|
6 |
+
# Custom
|
7 |
+
*.pt
|
8 |
+
data/local
|
9 |
+
# Byte-compiled / optimized / DLL files
|
10 |
+
__pycache__/
|
11 |
+
*.py[cod]
|
12 |
+
*$py.class
|
13 |
+
|
14 |
+
# C extensions
|
15 |
+
*.so
|
16 |
+
|
17 |
+
# Distribution / packaging
|
18 |
+
.Python
|
19 |
+
build/
|
20 |
+
develop-eggs/
|
21 |
+
dist/
|
22 |
+
downloads/
|
23 |
+
eggs/
|
24 |
+
.eggs/
|
25 |
+
lib/
|
26 |
+
lib64/
|
27 |
+
parts/
|
28 |
+
sdist/
|
29 |
+
var/
|
30 |
+
wheels/
|
31 |
+
pip-wheel-metadata/
|
32 |
+
share/python-wheels/
|
33 |
+
*.egg-info/
|
34 |
+
.installed.cfg
|
35 |
+
*.egg
|
36 |
+
MANIFEST
|
37 |
+
|
38 |
+
# PyInstaller
|
39 |
+
# Usually these files are written by a python script from a template
|
40 |
+
# before PyInstaller builds the exe, so as to inject date/other infos into it.
|
41 |
+
*.manifest
|
42 |
+
*.spec
|
43 |
+
|
44 |
+
# Installer logs
|
45 |
+
pip-log.txt
|
46 |
+
pip-delete-this-directory.txt
|
47 |
+
|
48 |
+
# Unit test / coverage reports
|
49 |
+
htmlcov/
|
50 |
+
.tox/
|
51 |
+
.nox/
|
52 |
+
.coverage
|
53 |
+
.coverage.*
|
54 |
+
.cache
|
55 |
+
nosetests.xml
|
56 |
+
coverage.xml
|
57 |
+
*.cover
|
58 |
+
*.py,cover
|
59 |
+
.hypothesis/
|
60 |
+
.pytest_cache/
|
61 |
+
|
62 |
+
# Translations
|
63 |
+
*.mo
|
64 |
+
*.pot
|
65 |
+
|
66 |
+
# Django stuff:
|
67 |
+
*.log
|
68 |
+
local_settings.py
|
69 |
+
db.sqlite3
|
70 |
+
db.sqlite3-journal
|
71 |
+
|
72 |
+
# Flask stuff:
|
73 |
+
instance/
|
74 |
+
.webassets-cache
|
75 |
+
|
76 |
+
# Scrapy stuff:
|
77 |
+
.scrapy
|
78 |
+
|
79 |
+
# Sphinx documentation
|
80 |
+
docs/_build/
|
81 |
+
|
82 |
+
# PyBuilder
|
83 |
+
target/
|
84 |
+
|
85 |
+
# Jupyter Notebook
|
86 |
+
.ipynb_checkpoints
|
87 |
+
|
88 |
+
# IPython
|
89 |
+
profile_default/
|
90 |
+
ipython_config.py
|
91 |
+
|
92 |
+
# pyenv
|
93 |
+
.python-version
|
94 |
+
|
95 |
+
# pipenv
|
96 |
+
# According to pypa/pipenv#598, it is recommended to include Pipfile.lock in version control.
|
97 |
+
# However, in case of collaboration, if having platform-specific dependencies or dependencies
|
98 |
+
# having no cross-platform support, pipenv may install dependencies that don't work, or not
|
99 |
+
# install all needed dependencies.
|
100 |
+
#Pipfile.lock
|
101 |
+
|
102 |
+
# PEP 582; used by e.g. github.com/David-OConnor/pyflow
|
103 |
+
__pypackages__/
|
104 |
+
|
105 |
+
# Celery stuff
|
106 |
+
celerybeat-schedule
|
107 |
+
celerybeat.pid
|
108 |
+
|
109 |
+
# SageMath parsed files
|
110 |
+
*.sage.py
|
111 |
+
|
112 |
+
# Environments
|
113 |
+
.env
|
114 |
+
.venv
|
115 |
+
env/
|
116 |
+
venv/
|
117 |
+
ENV/
|
118 |
+
env.bak/
|
119 |
+
venv.bak/
|
120 |
+
|
121 |
+
# Spyder project settings
|
122 |
+
.spyderproject
|
123 |
+
.spyproject
|
124 |
+
|
125 |
+
# Rope project settings
|
126 |
+
.ropeproject
|
127 |
+
|
128 |
+
# mkdocs documentation
|
129 |
+
/site
|
130 |
+
|
131 |
+
# mypy
|
132 |
+
.mypy_cache/
|
133 |
+
.dmypy.json
|
134 |
+
dmypy.json
|
135 |
+
|
136 |
+
# Pyre type checker
|
137 |
+
.pyre/
|
README.md
CHANGED
@@ -1,12 +1,6 @@
|
|
1 |
---
|
2 |
title: ViTExCo
|
3 |
-
|
4 |
-
colorFrom: gray
|
5 |
-
colorTo: green
|
6 |
sdk: gradio
|
7 |
sdk_version: 3.40.1
|
8 |
-
app_file: app.py
|
9 |
-
pinned: false
|
10 |
---
|
11 |
-
|
12 |
-
Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
|
|
|
1 |
---
|
2 |
title: ViTExCo
|
3 |
+
app_file: app.py
|
|
|
|
|
4 |
sdk: gradio
|
5 |
sdk_version: 3.40.1
|
|
|
|
|
6 |
---
|
|
|
|
UI.py
ADDED
@@ -0,0 +1,81 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import streamlit as st
|
2 |
+
from PIL import Image
|
3 |
+
import torchvision.transforms as transforms
|
4 |
+
from streamlit_image_comparison import image_comparison
|
5 |
+
import numpy as np
|
6 |
+
import torch
|
7 |
+
import torchvision
|
8 |
+
|
9 |
+
######################################### Utils ########################################
|
10 |
+
video_extensions = ["mp4"]
|
11 |
+
image_extensions = ["png", "jpg"]
|
12 |
+
|
13 |
+
|
14 |
+
def check_type(file_name: str):
|
15 |
+
for image_extension in image_extensions:
|
16 |
+
if file_name.endswith(image_extension):
|
17 |
+
return "image"
|
18 |
+
for video_extension in video_extensions:
|
19 |
+
if file_name.endswith(video_extension):
|
20 |
+
return "video"
|
21 |
+
return None
|
22 |
+
|
23 |
+
|
24 |
+
transform = transforms.Compose(
|
25 |
+
[transforms.Resize((256, 256)), transforms.ToTensor(), transforms.Normalize((0.485, 0.456, 0.406), (0.229, 0.224, 0.225))]
|
26 |
+
)
|
27 |
+
|
28 |
+
|
29 |
+
###################################### Load model ######################################
|
30 |
+
@st.cache_resource
|
31 |
+
def load_model():
|
32 |
+
model = torchvision.models.segmentation.deeplabv3_resnet101(pretrained=True)
|
33 |
+
model.eval()
|
34 |
+
return model
|
35 |
+
|
36 |
+
|
37 |
+
model = load_model()
|
38 |
+
########################################## UI ##########################################
|
39 |
+
st.title("Colorization")
|
40 |
+
|
41 |
+
uploaded_file = st.file_uploader("Upload grayscale image or video", type=image_extensions + video_extensions)
|
42 |
+
if uploaded_file:
|
43 |
+
# Image
|
44 |
+
if check_type(file_name=uploaded_file.name) == "image":
|
45 |
+
image = np.array(Image.open(uploaded_file), dtype=np.float32)
|
46 |
+
|
47 |
+
input_tensor = torchvision.transforms.functional.normalize(
|
48 |
+
torch.tensor(image).permute(2, 0, 1),
|
49 |
+
mean=[0.485, 0.456, 0.406],
|
50 |
+
std=[0.229, 0.224, 0.225],
|
51 |
+
).unsqueeze(0)
|
52 |
+
process_button = st.button("Process")
|
53 |
+
if process_button:
|
54 |
+
with st.spinner("Từ từ coi..."):
|
55 |
+
prediction = model(input_tensor)
|
56 |
+
segment = prediction["out"][0].permute(1, 2, 0)
|
57 |
+
segment = segment.detach().numpy()
|
58 |
+
|
59 |
+
st.image(segment)
|
60 |
+
st.image(image)
|
61 |
+
|
62 |
+
image_comparison(
|
63 |
+
img1=image,
|
64 |
+
img2=np.array(segment),
|
65 |
+
label1="Grayscale",
|
66 |
+
label2="Colorized",
|
67 |
+
make_responsive=True,
|
68 |
+
show_labels=True,
|
69 |
+
)
|
70 |
+
# Video
|
71 |
+
else:
|
72 |
+
# video = open(uploaded_file.name)
|
73 |
+
st.video("https://youtu.be/dQw4w9WgXcQ")
|
74 |
+
|
75 |
+
hide_menu_style = """
|
76 |
+
<style>
|
77 |
+
#MainMenu {visibility: hidden; }
|
78 |
+
footer {visibility: hidden;}
|
79 |
+
</style>
|
80 |
+
"""
|
81 |
+
st.markdown(hide_menu_style, unsafe_allow_html=True)
|
app.py
ADDED
@@ -0,0 +1,215 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import numpy as np
|
2 |
+
import shutil
|
3 |
+
import os
|
4 |
+
import argparse
|
5 |
+
import torch
|
6 |
+
import glob
|
7 |
+
from tqdm import tqdm
|
8 |
+
from PIL import Image
|
9 |
+
from collections import OrderedDict
|
10 |
+
from src.models.vit.config import load_config
|
11 |
+
import torchvision.transforms as transforms
|
12 |
+
import cv2
|
13 |
+
from skimage import io
|
14 |
+
|
15 |
+
from src.models.CNN.ColorVidNet import GeneralColorVidNet
|
16 |
+
from src.models.vit.embed import GeneralEmbedModel
|
17 |
+
from src.models.CNN.NonlocalNet import GeneralWarpNet
|
18 |
+
from src.models.CNN.FrameColor import frame_colorization
|
19 |
+
from src.utils import (
|
20 |
+
RGB2Lab,
|
21 |
+
ToTensor,
|
22 |
+
Normalize,
|
23 |
+
uncenter_l,
|
24 |
+
tensor_lab2rgb,
|
25 |
+
SquaredPadding,
|
26 |
+
UnpaddingSquare
|
27 |
+
)
|
28 |
+
|
29 |
+
import gradio as gr
|
30 |
+
|
31 |
+
def load_params(ckpt_file):
|
32 |
+
params = torch.load(ckpt_file, map_location=device)
|
33 |
+
new_params = []
|
34 |
+
for key, value in params.items():
|
35 |
+
new_params.append((key, value))
|
36 |
+
return OrderedDict(new_params)
|
37 |
+
|
38 |
+
def custom_transform(transforms, img):
|
39 |
+
for transform in transforms:
|
40 |
+
if isinstance(transform, SquaredPadding):
|
41 |
+
img,padding=transform(img, return_paddings=True)
|
42 |
+
else:
|
43 |
+
img = transform(img)
|
44 |
+
return img.to(device), padding
|
45 |
+
|
46 |
+
def save_frames(predicted_rgb, video_name, frame_name):
|
47 |
+
if predicted_rgb is not None:
|
48 |
+
predicted_rgb = np.clip(predicted_rgb, 0, 255).astype(np.uint8)
|
49 |
+
# frame_path_parts = frame_path.split(os.sep)
|
50 |
+
# if os.path.exists(os.path.join(OUTPUT_RESULT_PATH, frame_path_parts[-2])):
|
51 |
+
# shutil.rmtree(os.path.join(OUTPUT_RESULT_PATH, frame_path_parts[-2]))
|
52 |
+
# os.makedirs(os.path.join(OUTPUT_RESULT_PATH, frame_path_parts[-2]), exist_ok=True)
|
53 |
+
predicted_rgb = np.transpose(predicted_rgb, (1,2,0))
|
54 |
+
pil_img = Image.fromarray(predicted_rgb)
|
55 |
+
pil_img.save(os.path.join(OUTPUT_RESULT_PATH, video_name, frame_name))
|
56 |
+
|
57 |
+
def extract_frames_from_video(video_path):
|
58 |
+
cap = cv2.VideoCapture(video_path)
|
59 |
+
fps = cap.get(cv2.CAP_PROP_FPS)
|
60 |
+
|
61 |
+
# remove if exists folder
|
62 |
+
output_frames_path = os.path.join(INPUT_VIDEO_FRAMES_PATH, os.path.basename(video_path))
|
63 |
+
if os.path.exists(output_frames_path):
|
64 |
+
shutil.rmtree(output_frames_path)
|
65 |
+
|
66 |
+
# make new folder
|
67 |
+
os.makedirs(output_frames_path)
|
68 |
+
|
69 |
+
currentframe = 0
|
70 |
+
frame_path_list = []
|
71 |
+
while(True):
|
72 |
+
|
73 |
+
# reading from frame
|
74 |
+
ret,frame = cap.read()
|
75 |
+
|
76 |
+
if ret:
|
77 |
+
name = os.path.join(output_frames_path, f'{currentframe:09d}.jpg')
|
78 |
+
frame_path_list.append(name)
|
79 |
+
cv2.imwrite(name, frame)
|
80 |
+
currentframe += 1
|
81 |
+
else:
|
82 |
+
break
|
83 |
+
|
84 |
+
cap.release()
|
85 |
+
cv2.destroyAllWindows()
|
86 |
+
|
87 |
+
return frame_path_list, fps
|
88 |
+
|
89 |
+
def combine_frames_from_folder(frames_list_path, fps = 30):
|
90 |
+
frames_list = glob.glob(f'{frames_list_path}/*.jpg')
|
91 |
+
frames_list.sort()
|
92 |
+
|
93 |
+
sample_shape = cv2.imread(frames_list[0]).shape
|
94 |
+
|
95 |
+
output_video_path = os.path.join(frames_list_path, 'output_video.mp4')
|
96 |
+
out = cv2.VideoWriter(output_video_path, cv2.VideoWriter_fourcc(*'mp4v'), fps, (sample_shape[1], sample_shape[0]))
|
97 |
+
for filename in frames_list:
|
98 |
+
img = cv2.imread(filename)
|
99 |
+
out.write(img)
|
100 |
+
|
101 |
+
out.release()
|
102 |
+
return output_video_path
|
103 |
+
|
104 |
+
|
105 |
+
def upscale_image(I_current_rgb, I_current_ab_predict):
|
106 |
+
H, W = I_current_rgb.size
|
107 |
+
high_lab_transforms = [
|
108 |
+
SquaredPadding(target_size=max(H,W)),
|
109 |
+
RGB2Lab(),
|
110 |
+
ToTensor(),
|
111 |
+
Normalize()
|
112 |
+
]
|
113 |
+
# current_frame_pil_rgb = Image.fromarray(np.clip(I_current_rgb.squeeze(0).permute(1,2,0).cpu().numpy() * 255, 0, 255).astype('uint8'))
|
114 |
+
high_lab_current, paddings = custom_transform(high_lab_transforms, I_current_rgb)
|
115 |
+
high_lab_current = torch.unsqueeze(high_lab_current,dim=0).to(device)
|
116 |
+
high_l_current = high_lab_current[:, 0:1, :, :]
|
117 |
+
high_ab_current = high_lab_current[:, 1:3, :, :]
|
118 |
+
upsampler = torch.nn.Upsample(scale_factor=max(H,W)/224,mode="bilinear")
|
119 |
+
high_ab_predict = upsampler(I_current_ab_predict)
|
120 |
+
I_predict_rgb = tensor_lab2rgb(torch.cat((uncenter_l(high_l_current), high_ab_predict), dim=1))
|
121 |
+
upadded = UnpaddingSquare()
|
122 |
+
I_predict_rgb = upadded(I_predict_rgb, paddings)
|
123 |
+
return I_predict_rgb
|
124 |
+
|
125 |
+
def colorize_video(video_path, ref_np):
|
126 |
+
frames_list, fps = extract_frames_from_video(video_path)
|
127 |
+
|
128 |
+
frame_ref = Image.fromarray(ref_np).convert("RGB")
|
129 |
+
I_last_lab_predict = None
|
130 |
+
IB_lab, IB_paddings = custom_transform(transforms, frame_ref)
|
131 |
+
IB_lab = IB_lab.unsqueeze(0).to(device)
|
132 |
+
IB_l = IB_lab[:, 0:1, :, :]
|
133 |
+
IB_ab = IB_lab[:, 1:3, :, :]
|
134 |
+
|
135 |
+
with torch.no_grad():
|
136 |
+
I_reference_lab = IB_lab
|
137 |
+
I_reference_l = I_reference_lab[:, 0:1, :, :]
|
138 |
+
I_reference_ab = I_reference_lab[:, 1:3, :, :]
|
139 |
+
I_reference_rgb = tensor_lab2rgb(torch.cat((uncenter_l(I_reference_l), I_reference_ab), dim=1)).to(device)
|
140 |
+
features_B = embed_net(I_reference_rgb)
|
141 |
+
|
142 |
+
video_path_parts = frames_list[0].split(os.sep)
|
143 |
+
|
144 |
+
if os.path.exists(os.path.join(OUTPUT_RESULT_PATH, video_path_parts[-2])):
|
145 |
+
shutil.rmtree(os.path.join(OUTPUT_RESULT_PATH, video_path_parts[-2]))
|
146 |
+
os.makedirs(os.path.join(OUTPUT_RESULT_PATH, video_path_parts[-2]), exist_ok=True)
|
147 |
+
|
148 |
+
for frame_path in tqdm(frames_list):
|
149 |
+
curr_frame = Image.open(frame_path).convert("RGB")
|
150 |
+
IA_lab, IA_paddings = custom_transform(transforms, curr_frame)
|
151 |
+
IA_lab = IA_lab.unsqueeze(0).to(device)
|
152 |
+
IA_l = IA_lab[:, 0:1, :, :]
|
153 |
+
IA_ab = IA_lab[:, 1:3, :, :]
|
154 |
+
|
155 |
+
if I_last_lab_predict is None:
|
156 |
+
I_last_lab_predict = torch.zeros_like(IA_lab).to(device)
|
157 |
+
|
158 |
+
with torch.no_grad():
|
159 |
+
I_current_lab = IA_lab
|
160 |
+
I_current_ab_predict, _ = frame_colorization(
|
161 |
+
IA_l,
|
162 |
+
I_reference_lab,
|
163 |
+
I_last_lab_predict,
|
164 |
+
features_B,
|
165 |
+
embed_net,
|
166 |
+
nonlocal_net,
|
167 |
+
colornet,
|
168 |
+
luminance_noise=0,
|
169 |
+
temperature=1e-10,
|
170 |
+
joint_training=False
|
171 |
+
)
|
172 |
+
I_last_lab_predict = torch.cat((IA_l, I_current_ab_predict), dim=1)
|
173 |
+
|
174 |
+
# IA_predict_rgb = tensor_lab2rgb(torch.cat((uncenter_l(IA_l), I_current_ab_predict), dim=1))
|
175 |
+
IA_predict_rgb = upscale_image(curr_frame, I_current_ab_predict)
|
176 |
+
#IA_predict_rgb = torch.nn.functional.upsample_bilinear(IA_predict_rgb, scale_factor=2)
|
177 |
+
save_frames(IA_predict_rgb.squeeze(0).cpu().numpy() * 255, video_path_parts[-2], os.path.basename(frame_path))
|
178 |
+
return combine_frames_from_folder(os.path.join(OUTPUT_RESULT_PATH, video_path_parts[-2]), fps)
|
179 |
+
|
180 |
+
if __name__ == '__main__':
|
181 |
+
# Init global variables
|
182 |
+
device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
|
183 |
+
INPUT_VIDEO_FRAMES_PATH = 'inputs'
|
184 |
+
OUTPUT_RESULT_PATH = 'outputs'
|
185 |
+
weight_path = 'checkpoints'
|
186 |
+
|
187 |
+
embed_net=GeneralEmbedModel(pretrained_model="swin-tiny", device=device).to(device)
|
188 |
+
nonlocal_net = GeneralWarpNet(feature_channel=128).to(device)
|
189 |
+
colornet=GeneralColorVidNet(7).to(device)
|
190 |
+
|
191 |
+
embed_net.eval()
|
192 |
+
nonlocal_net.eval()
|
193 |
+
colornet.eval()
|
194 |
+
|
195 |
+
# Load weights
|
196 |
+
# embed_net_params = load_params(os.path.join(weight_path, "embed_net.pth"))
|
197 |
+
nonlocal_net_params = load_params(os.path.join(weight_path, "nonlocal_net.pth"))
|
198 |
+
colornet_params = load_params(os.path.join(weight_path, "colornet.pth"))
|
199 |
+
|
200 |
+
# embed_net.load_state_dict(embed_net_params, strict=True)
|
201 |
+
nonlocal_net.load_state_dict(nonlocal_net_params, strict=True)
|
202 |
+
colornet.load_state_dict(colornet_params, strict=True)
|
203 |
+
|
204 |
+
transforms = [SquaredPadding(target_size=224),
|
205 |
+
RGB2Lab(),
|
206 |
+
ToTensor(),
|
207 |
+
Normalize()]
|
208 |
+
|
209 |
+
examples = [[vid, ref] for vid, ref in zip(sorted(glob.glob('examples/*/*.mp4')), sorted(glob.glob('examples/*/*.jpg')))]
|
210 |
+
demo = gr.Interface(colorize_video,
|
211 |
+
inputs=[gr.Video(), gr.Image()],
|
212 |
+
outputs="playable_video",
|
213 |
+
examples=examples,
|
214 |
+
cache_examples=True)
|
215 |
+
demo.launch()
|
checkpoints/colornet.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:5257ae325e292cd5fb2eff47095e1c4e4815455bd5fb6dc5ed2ee2b923172875
|
3 |
+
size 131239411
|
checkpoints/embed_net.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:fc711755a75c43025dabe9407cbd11d164eaa9e21f26430d0c16c7493410d902
|
3 |
+
size 110352261
|
checkpoints/nonlocal_net.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:b94c6990f20088bc3cc3fe0b29a6d52e6e746b915c506f0cd349fc6ad6197e72
|
3 |
+
size 73189765
|
cmd.txt
ADDED
@@ -0,0 +1,21 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
python train.py --video_data_root_list datasets/images/images \
|
2 |
+
--flow_data_root_list datasets/flow_fp16/flow_fp16 \
|
3 |
+
--mask_data_root_list datasets/pgm/pgm \
|
4 |
+
--data_root_imagenet datasets/imgnet \
|
5 |
+
--annotation_file_path datasets/final_annot.csv \
|
6 |
+
--imagenet_pairs_file datasets/pairs.txt \
|
7 |
+
--gpu_ids 0 \
|
8 |
+
--workers 12 \
|
9 |
+
--batch_size 2 \
|
10 |
+
--real_reference_probability 0.99 \
|
11 |
+
--weight_contextual 1 \
|
12 |
+
--weight_perceptual 0.1 \
|
13 |
+
--weight_smoothness 5 \
|
14 |
+
--weight_gan 0.9 \
|
15 |
+
--weight_consistent 0.1 \
|
16 |
+
--use_wandb True \
|
17 |
+
--wandb_token "f05d31e6b15339b1cfc5ee1c77fe51f66fc3ea9e" \
|
18 |
+
--wandb_name "vit_tiny_patch16_384_nofeat" \
|
19 |
+
--checkpoint_step 500 \
|
20 |
+
--epoch_train_discriminator 3 \
|
21 |
+
--epoch 20
|
cmd_ddp.txt
ADDED
@@ -0,0 +1,20 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
!torchrun --nnodes=1 --nproc_per_node=2 train_ddp.py --video_data_root_list $video_data_root_list \
|
2 |
+
--flow_data_root_list $flow_data_root_list \
|
3 |
+
--mask_data_root_list $mask_data_root_list \
|
4 |
+
--data_root_imagenet $data_root_imagenet \
|
5 |
+
--annotation_file_path $annotation_file_path \
|
6 |
+
--imagenet_pairs_file $imagenet_pairs_file \
|
7 |
+
--gpu_ids "0,1" \
|
8 |
+
--workers 2 \
|
9 |
+
--batch_size 2 \
|
10 |
+
--real_reference_probability 0.99 \
|
11 |
+
--weight_contextual 1 \
|
12 |
+
--weight_perceptual 0.1 \
|
13 |
+
--weight_smoothness 5 \
|
14 |
+
--weight_gan 0.9 \
|
15 |
+
--weight_consistent 0.1 \
|
16 |
+
--wandb_token "165e7148081f263b423722115e2ad40fa5339ecf" \
|
17 |
+
--wandb_name "vit_tiny_patch16_384_nofeat" \
|
18 |
+
--checkpoint_step 2000 \
|
19 |
+
--epoch_train_discriminator 2 \
|
20 |
+
--epoch 10
|
docs/.gitignore
ADDED
File without changes
|
environment.yml
ADDED
File without changes
|
examples.zip
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:bd4531bd3abdec6df90efb0d19fadd54284bdc70d5edfff19752a205159eb4db
|
3 |
+
size 6955837
|
examples/bear/ref.jpg
ADDED
![]() |
examples/bear/video.mp4
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:cb4cec5064873a4616f78bdb653830683a4842b2a5cfd0665b395cff4d120d04
|
3 |
+
size 1263445
|
examples/boat/ref.jpg
ADDED
![]() |
examples/boat/video.mp4
ADDED
Binary file (853 kB). View file
|
|
examples/cows/ref.jpg
ADDED
![]() |
examples/cows/video.mp4
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:1ac08603d719cd7a8d71fac76c9318d3e8f1e516e9b3c2a06323a0e4e78f6410
|
3 |
+
size 2745681
|
examples/flamingo/ref.jpg
ADDED
![]() |
examples/flamingo/video.mp4
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:5a103fd4991a00e419e5236b885fe9d220704ba0a6ac794c87aaa3f62a4f1561
|
3 |
+
size 1239570
|
gradio_cached_examples/13/log.csv
ADDED
@@ -0,0 +1,5 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
output,flag,username,timestamp
|
2 |
+
/content/ViTExCo/gradio_cached_examples/13/output/003c3114319372a78bf2f812ebaf0041afa280fb/output_video.mp4,,,2023-08-15 09:45:37.897615
|
3 |
+
/content/ViTExCo/gradio_cached_examples/13/output/e6d6153dedeb9fec586b3241311cc49dbc17bc85/output_video.mp4,,,2023-08-15 09:46:01.048997
|
4 |
+
/content/ViTExCo/gradio_cached_examples/13/output/7969adca8ae38cb3b38ff8e7bb54688d942c7bc8/output_video.mp4,,,2023-08-15 09:46:34.503322
|
5 |
+
/content/ViTExCo/gradio_cached_examples/13/output/74c76e483235b7e80665e32d7fcdcc3da2be7644/output_video.mp4,,,2023-08-15 09:46:58.088903
|
gradio_cached_examples/13/output/003c3114319372a78bf2f812ebaf0041afa280fb/output_video.mp4
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:b5ab666998e14fb00281a90f8801753eca001a432641ae2770007a8336b4c64e
|
3 |
+
size 1213824
|
gradio_cached_examples/13/output/74c76e483235b7e80665e32d7fcdcc3da2be7644/output_video.mp4
ADDED
Binary file (914 kB). View file
|
|
gradio_cached_examples/13/output/7969adca8ae38cb3b38ff8e7bb54688d942c7bc8/output_video.mp4
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:7c367dab34e596f7f0fed34c7e2384525de2ba1824b410d0770bdbd17bc9e72a
|
3 |
+
size 1793060
|
gradio_cached_examples/13/output/e6d6153dedeb9fec586b3241311cc49dbc17bc85/output_video.mp4
ADDED
Binary file (673 kB). View file
|
|
inputs/video.mp4/000000000.jpg
ADDED
![]() |
inputs/video.mp4/000000001.jpg
ADDED
![]() |
inputs/video.mp4/000000002.jpg
ADDED
![]() |
inputs/video.mp4/000000003.jpg
ADDED
![]() |
inputs/video.mp4/000000004.jpg
ADDED
![]() |
inputs/video.mp4/000000005.jpg
ADDED
![]() |
inputs/video.mp4/000000006.jpg
ADDED
![]() |
inputs/video.mp4/000000007.jpg
ADDED
![]() |
inputs/video.mp4/000000008.jpg
ADDED
![]() |
inputs/video.mp4/000000009.jpg
ADDED
![]() |
inputs/video.mp4/000000010.jpg
ADDED
![]() |
inputs/video.mp4/000000011.jpg
ADDED
![]() |
inputs/video.mp4/000000012.jpg
ADDED
![]() |
inputs/video.mp4/000000013.jpg
ADDED
![]() |
inputs/video.mp4/000000014.jpg
ADDED
![]() |
inputs/video.mp4/000000015.jpg
ADDED
![]() |
inputs/video.mp4/000000016.jpg
ADDED
![]() |
inputs/video.mp4/000000017.jpg
ADDED
![]() |
inputs/video.mp4/000000018.jpg
ADDED
![]() |
inputs/video.mp4/000000019.jpg
ADDED
![]() |
inputs/video.mp4/000000020.jpg
ADDED
![]() |
inputs/video.mp4/000000021.jpg
ADDED
![]() |
inputs/video.mp4/000000022.jpg
ADDED
![]() |
inputs/video.mp4/000000023.jpg
ADDED
![]() |