Changelog

0.4.0

This is the first release of Torch Tweak. Torch Tweak is focused on Intel GPU (XPU) support. This release introduces i.a. new backend, tighter dependency and security updates, code cleanups, and new continuous integration system.

This is a pre-release (pre-1.0) version and not yet production-quality.

Please note, this project was forked from repository https://github.com/ai-dynamo/aitune, from version 0.3.0. To see its changelog go to the upstream repository.

Details: - feat: add XPU support with Intel GPU-focused installation requirements and device handling throughout the toolkit - feat: add OpenVINO backend with tests - feat: improve tuning flow for XPU workloads, including hybrid XPU benchmark support - fix: improve robustness for safety-critical file handling and I/O operations - fix: update dependency constraints and security patches for vulnerable packages and compatibility issues - misc: streamline docs, CI, and environment setup for XPU-based development and deployment - misc: change default save extension from .ait to .tt - misc: unblock latest Python 3.13+