Extend platform compatibility to ARM based devices (Apple, Ampere)

Łukasz Wałejkorequested to merge
arm into main
  • Created pyproject.toml for project metadata and dependencies.
  • Added setup.sh for environment setup and dependency installation based on platform.
  • Introduced test_platforms.sh to validate platform configurations and dependencies.
  • Implemented test_platform.py to check CUDA, MPS, and CPU tensor operations.
  • Updated Dockerfiles for building images with appropriate dependencies.

Summary by CodeRabbit

  • New Features

    • Centralized prediction models configuration for easy model selection.
    • New automated Docker build script supporting CPU/CUDA variants.

  • Improvements

    • Unified device selection for AI models (auto/CPU/CUDA/MPS) and reduced verbose model output.
    • Updated Docker images and container setup for improved environment isolation.
    • CI refactor with reusable build template and improved image tagging/copying.

  • Bug Fixes

    • Suppressed noisy deprecation warnings; pinned/added key dependencies.

  • Chores

    • Updated ignore, editor settings, docs, and compose formatting.
Edited by OSCF CodeRabbit

Merge request reports