Release v1.1.0-patched (Patched)

Qualcomm AI Stack SDK - Patched Version v1.1.0-patched

This release contains critical memory corruption fixes for the Qualcomm AI Stack model converter tools.

Key Patches Included:

• Python-based Shape Re-inference: Robust topological pass to fix C++ memory corruption.
• Memory Corruption Prevention: Python-side caching of critical tensor metadata.
• Op-Specific Robustness: Fixed Conv2D, MatMul, and BatchNorm transformations.
• Stability Fixes: Removed unstable C++ calls in the graph optimization layer.

Support the Developer

If these fixes helped your project, consider making a donation:
Donate via PayPal