https://github.com/ap--/nvjpeg2k-numcodecs
nvJPEG2000 numcodecs playground
https://github.com/ap--/nvjpeg2k-numcodecs
Last synced: 9 months ago
JSON representation
nvJPEG2000 numcodecs playground
- Host: GitHub
- URL: https://github.com/ap--/nvjpeg2k-numcodecs
- Owner: ap--
- License: mit
- Created: 2022-10-25T19:43:05.000Z (over 3 years ago)
- Default Branch: main
- Last Pushed: 2023-03-22T22:04:55.000Z (about 3 years ago)
- Last Synced: 2025-06-04T05:50:51.219Z (about 1 year ago)
- Language: Cython
- Size: 41 KB
- Stars: 4
- Watchers: 1
- Forks: 1
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- Changelog: CHANGELOG.md
- License: LICENSE
Awesome Lists containing this project
README
# nvjpeg2k_numcodecs
This is a playground for getting nvjpeg2k decoding into numcodecs to
benchmark if tile loading for deeplearning can be significantly sped
up when decoding directly on the GPU.
## :construction: Notes :construction:
Currently, this is a bit of a playground for testing feasibility and
establishing good benchmarks. But feel free to open issues in the
issue tracker in case you would like to help and or contribute :heart:
## Documentation
### Installation
Currently no wheels or conda packages.
To get started you need cuda and a cupy installation.
Then clone the repo and run:
```shell
cd vendor
./get_nvjpeg2k.sh
```
after that you should be able to run the following in the repository root:
```shell
pip install .
```
## Contributing Guidelines
Please use `pre-commit` unless it interferes with your workflow too much.
I'm happy to get contributions, so don't let linting be in the way of that.
## Acknowledgements
Build with love by Andreas Poehlmann.
`nvjpeg2k_numcodecs`: copyright 2022 Andreas Poehlmann, licensed under [MIT](https://github.com/ap--/nvjpeg2k_numcodecs/blob/master/LICENSE)