Video_Colorization_CodeBase_CVPR23_NTIRE
  • 模型资讯
  • 模型资料

NTIRE 2023 Video Colorization Model and Dataset

This project provides the baseline model and evaluation code for track1 and track2 for CVPR 2023 NTIRE workshop Video Colorization Challenge.

Installation

conda create -n video_colorization python=3.7
conda activate video_colorization

pip install torch==1.12.1+cu116 torchvision==0.13.1+cu116 torchaudio==0.12.1 --extra-index-url https://download.pytorch.org/whl/cu116

git clone https://github.com/piddnad/CVPR2023_NTIRE_Video_Colorization.git

cd CVPR2023_NTIRE_Video_Colorization
pip install -r requirements/tests.txt
pip install -r requirements/framework.txt
pip install -r requirements/cv.txt

Download Dataset (Optional)

You can Run the code below to download the validation set:

from modelscope.msdatasets import MsDataset
from modelscope.utils.constant import DownloadMode


# Set dataset download path
cache_dir = './datasets' 

# Download validation set
val_set = MsDataset.load('ntire23_video_colorization', namespace='damo', subset_name='val_frames', split='validation', cache_dir=cache_dir, download_mode=DownloadMode.FORCE_REDOWNLOAD)
print(next(iter(val_set)))

Baseline Evaluation on Validation Set

This step will automatically download the validation set.

cd CVPR2023_NTIRE_Video_Colorization
CUDA_VISIBLE_DEVICES=0  PYTHONPATH=. python ntire23_scripts/baseline_evaluation.py

# Then you might get output similar to:
# FID evaluation time: xxxx
# CDC evaluation time: xxxx
# Total evaluation time: xxxx
# FID: 47.15574537543114, CDC: 0.003475072230336491

Evaluation on Your Results

First modify the res_dir in user_result_evaluation.py, and then run:

python ntire23_scripts/user_result_evaluation.py

Clone with HTTP

 git clone https://www.modelscope.cn/damo/CVPR2023_NTIRE_Video_Colorization.git