Mia
@MiaAI_lab
Building with AI & LLMs | Insights, recipes, tools & honest experiments
加入 July 2022
364 正在關注 12.3K 粉絲
You can now run GLM-5.2 on 3x @NVIDIAAI DGX Sparks at home 🤯
• 248k context
• 15-20 tok/s (content-dependent)
• MTP for max context / DSpark optional
• NVFP4+AQLM hybrid
Still a work in progress — will keep improving over time.
→
顯示更多
