3 Commits

Include renames

Author	SHA1	Message	Date
Jamie Pine	f96eae2567	fix: address PR review feedback from CodeRabbit ... - Upgrade softprops/action-gh-release@v1 to @v2 (Node 16 EOL) - Fail-fast on checksum fetch failure instead of extracting unverified archives - Abort packaging if no NVIDIA files found (prevents empty cuda-libs archive) - Fix nvidia/ path detection bug (list membership vs substring check) - Fix justfile Copy-Item nesting (copy contents, not the directory itself)	2026-03-17 06:53:54 -07:00
Jamie Pine	88be097b62	fix: update package_cuda.py for PyInstaller 6.18 layout and remove split_binary.py ... - Fix is_nvidia_file() to match NVIDIA DLLs in _internal/torch/lib/ (PyInstaller 6.18 + torch 2.10 no longer uses nvidia/ subdirectories) - Remove deprecated split_binary.py (both archives are under 2GB) - Update torch_compat range to >=2.6.0,<2.11.0 - Update build docs for new dual-archive packaging flow	2026-03-17 04:57:05 -07:00
James Pine	564d787927	feat: split CUDA backend into independently versioned server + libs archives ... Switch CUDA builds from PyInstaller --onefile to --onedir and split the output into two separately versioned archives: 1. Server core (~200-400MB) — versioned with the app, redownloaded on every app update 2. CUDA libs (~2GB) — versioned independently (cu126-v1), only redownloaded when the CUDA toolkit or torch version changes This eliminates the ~2.4GB full redownload on every version bump. After initial setup, most app updates only need ~200-400MB. Closes #297	2026-03-17 04:04:17 -07:00