v1.10.1 — Forge Neo integration, update checker, GPU layers fix

Two variants attached once the build finishes:

• AI_Launcher-cuda-v1.10.1.zip — for NVIDIA GPU users, Image Gen runs at real speed
• AI_Launcher-cpu-v1.10.1.zip — everything else; Image Gen still works, just slow without CUDA

Extract, copy config.example.json to config.json next to the exe (or launch once and use the Settings GUI — it prompts you if config.json is missing), fill in your paths, run AI_Launcher.exe. No Python install required.

Since v1.9.3:

• Added Forge Neo integration for Image Gen
• Added an in-app update checker
• Added a Restart App button, fixed it racing its own single-instance mutex
• Fixed the GPU Layers setting silently corrupting "Auto" (-1) to 0 on Save, which could force a model fully onto CPU without warning

Build is running now — zips will appear on this release once the workflow finishes (check the Actions tab for progress).

v1.9.3 — Windows packaging

First packaged release. Two variants attached:

• AI_Launcher-cuda-v1.9.3.zip — for NVIDIA GPU users, Image Gen runs at real speed
• AI_Launcher-cpu-v1.9.3.zip — everything else; Image Gen still works, just slow without CUDA

Extract, copy config.example.json to config.json next to the exe (or launch once and use the Settings GUI — it prompts you if config.json is missing), fill in your paths, run AI_Launcher.exe. No Python install required.

Also in this release: fixed a path-resolution bug that would have broken config/prefs/log persistence under the packaged build, added a Settings field for sdxl.dir, and added a Suggested Models section to the README.

Build is running now — zips will appear on this release once the workflow finishes (check the Actions tab for progress).