Releases: DIAGNijmegen/rse-grand-challenge-forge
Releases · DIAGNijmegen/rse-grand-challenge-forge
Release v0.6.1
What's Changed
- Use example values for JSON interfaces in examples by @chrisvanrun in #65
- Fix permissions on bash scripts by @chrisvanrun in #62
- Use challenge url in pack README by @chrisvanrun in #63
- Ensures .tif also get read in example inteference.py by @chrisvanrun in #66
Full Changelog: v0.6.0...v0.6.1
v0.6.0
Small release:
- Add forging of algorithm templates.
- Split CLI commands into 'pack' for challenge packs and 'algorithm' for algorithm templates.
v0.5.0
A large release, with a long list of improvement and changes:
- Fixed interoperability issues when running challenge pack scripts on macOS.
- Fixed several Docker deprecation warnings.
- Added examples for file inputs.
- Reworked multithreading in example evaluation methods:
- Used concurrent.futures.ProcessPoolExecutor.
- Dynamically limited the number of workers in evaluation methods (via multiprocessing.cpu_count()).
- Introduced a controller thread.
- Silenced Docker volume shell calls.
- Mounted a dummy /tmp to test scripts, preventing final containers that users might build.
- Challenge packs now include a proper .gitignore.
- Added a save.sh script to Docker-based examples to demonstrate how to export a Docker image.
- Ensured Docker file-mounted inputs are read-only.
- Dist challenge packs under the Apache License 2.0.
Release v0.4.0
Stop using '/tmp' in example Dockerfiles to contain non-transient files.
Release v0.3.0
Replace "input"/"output" in the context JSON input with "algorithm_input"/"algorithm_output".
Release v0.2.2
- Various improvements to spelling and wording in the challenge packs
- Report grand-challenge-forge version in generated pack README
- Use challenge-level archives instead of phase-level archives for the pack README
- Add support for reading and writing TIFF files
- Disable network in test runs
Release v0.2.1
- Various improvements to spelling and comments in the challenge packs
- Fixed passing
'is_image'
,'is_json'
to the prediction.json example of the evaluation method - Added a generic
read_file()
/write_file()
to the evaluation.py and inference.py of the algorithm and evaluation examples, respectively
Release v0.2.0
Adds challenge packs, consisting of:
- example upload algorithm
- example submission algorithm
- example evaluation method