Skip to content

Release Notes

0.1.14 [2024-01-12]

Fixes:

  • Fixed model path configuration in CI to use traced_ts_model.pt instead of traced_ts_model
  • Corrected MONAI package command to properly reference model file with full path ${MONAI_MAP_PATH}/${MONAI_MODEL_PATH}

0.1.13 [2023-12-14]

Maintenance:

  • Updated monai-deploy-app-sdk from 0.5.0 to 0.5.1

0.1.12 [2023-12-09]

Fixes:

  • Fixed python interpreter path in command from /usr/bin/python to /opt/conda/bin/python

0.1.11 [2023-11-21]

Fixes:

  • Downgraded base image from nvcr.io/nvidia/pytorch:23.08-py3 to nvcr.io/nvidia/pytorch:22.08-py3 for compatibility
  • Downgraded python requirement from ^3.10 to ^3.8 for broader compatibility

0.1.10 [2023-10-25]

Enhancements:

  • Added support for processing ZIP archive inputs containing DICOM files

Fixes:

  • Fixed Docker image publishing to use flywheel/ registry consistently

Maintenance:

  • Updated torch from >=2.0.0, !=2.0.1 to =2.0.0
  • Updated monai-deploy-app-sdk from 0.5.1 to 0.5.0
  • Added CI configuration tags for large Docker builds
  • Updated base image reference in documentation from pytorch:22.08-py3 to pytorch:23.08-py3
  • Removed version parameter from ModelInfo initialization
  • Simplified CI pipeline Docker image variable handling

0.1.8 [2023-09-14]

Enhancements:

  • Updated python from 3.8 to 3.10
  • Added torch>=2.0.0, !=2.0.1 dependency for improved PyTorch compatibility

Maintenance:

  • Updated base image from nvcr.io/nvidia/pytorch:22.08-py3 to nvcr.io/nvidia/pytorch:23.08-py3
  • Updated monai-deploy-app-sdk from 0.5.0 to 0.5.1

0.1.7 [2023-03-16]

Enhancements:

  • Added GPU support with NVIDIA CUDA environment configuration including NVIDIA_VISIBLE_DEVICES, NVIDIA_DRIVER_CAPABILITIES, and NVIDIA_REQUIRE_CUDA environment variables

0.1.6 [2023-03-15]

Documentation:

  • Added source URL reference to MONAI Model Zoo
  • Added project repository URL to manifest

0.1.5 [2023-02-17]

Enhancements:

  • Added breast density classification gear for mammogram images using MONAI Model Zoo
  • Added support for BI-RADS category classification (A-D) on DICOM mammogram images
  • Added file metadata updates with classification results in Flywheel platform
  • Added optional JSON export of model results and job details via save-json config flag

Documentation:

  • Added comprehensive README with gear overview, usage instructions, and workflow diagrams
  • Added CONTRIBUTING guide with development setup, dependency management, and release process
  • Added implementation details documentation for MONAI Application Package (MAP) build

Maintenance:

  • Added CI/CD pipeline configuration with pre-commit hooks, Docker publishing, and release automation
  • Added flywheel-gear-toolkit 0.6.8 for Flywheel integration
  • Added monai 1.1.0 and monai-deploy-app-sdk 0.5.0 for model deployment
  • Added pydicom 2.3.1 and highdicom 0.20.0 for DICOM processing
  • Added pre-commit hooks for black, markdownlint, and manifest validation