Back to home
LocalAI
2 articles tagged with this topic
MicrosoftVibeVoice
Microsoft VibeVoice Runs Without Python — AI De-Pythonization Hits Speech
Microsoft VibeVoice ported to pure C++ — no Python for inference. AI's de-Pythonization trend expands from text to voice, lowering enterprise voice AI
May 52 min read
Qwen3.5LocalAI
Qwen 3.5 35B Benchmarks: Vulkan vs ROCm on AMD Strix Halo
Vulkan wins token generation (~57.5 t/s) while ROCm leads prompt processing (~1052 t/s) on AMD Ryzen AI MAX+ 395.
Apr 83 min read