Prince Canuma
@Prince_Canuma
加入 July 2012
1.3K 正在關注 22.8K 粉絲
LFM2.5 DSpark by @liquidai is coming to mlx-vlm in v0.6.16 ⚡️
Exact speculative decoding on M5 Max, delivering up to 3.7× faster generation with zero output drift.
On-device speed, zero output drift.
Benchmarks below 🧵
顯示更多
