Skip to content

Latest commit

 

History

History
93 lines (59 loc) · 2.22 KB

README.md

File metadata and controls

93 lines (59 loc) · 2.22 KB
Raw

Readme for TCVC-Modal

TCVC-Modal is an implementation of the TCVC method for image colorization using PyTorch.

Todo

  • Add batch processing to optimize inference.

Installation

Option 1: Install natively

To install TCVC-Modal natively, follow these steps:

  1. Clone the repository to your machine using the following command:

    git clone https://github.com/cudanexus/tcvc
cd tcvc

  2. Run the installation script as root:

    sudo bash install.sh

  3. Install the required Python packages using pip:

    pip install -r requirements.txt

  4. Install the required packages for channelnorm_package, correlation_package, and resample2d_package:

    cd codes/models/archs/networks/channelnorm_package/ && python setup.py develop
cd codes/models/archs/networks/correlation_package/ && python setup.py develop
cd codes/models/archs/networks/resample2d_package/ && python setup.py develop
cd ../../../../

  5. Download the colorization backbone model and the pretrained models:

    wget "https://tcvc.s3.amazonaws.com/TCVC_IDC.zip"
wget "https://tcvc.s3.amazonaws.com/pretrained_models.zip"

  6. Extract the downloaded models and paste them in the experiments folder.

  7. Run the colorization test:

    python test_TCVC_onesampling_noGT.py

Option 2: Use Docker

To use TCVC-Modal with Docker, follow these steps:

  1. Clone the repository to your machine using the following command:

    git clone https://github.com/cudanexus/tcvc
cd tcvc

  2. Build the Docker image:

    docker build -t tcvc .

  3. Run the Docker container:

    docker run --gpus all --net host --rm -it --name tcvc tcvc

  4. Follow steps 5-7 above to run the colorization test inside the Docker container.

License

The code is available under the MIT License. Please see the LICENSE file for more information.

Acknowledgments

This implementation is based on the following paper:

  • Cheng, Zezhou, Qingxiong Yang, and Bin Sheng. "Deep colorization." Proceedings of the IEEE International Conference on Computer Vision. 2015.

Thanks to the authors for making their work available to the research community.