SOTA ternary-packed versions of 1.58-bit LLMs for efficient on-device inference with vlut.cpp.
Xiangyu Li
XXXXyu
AI & ML interests
On-device AI and physical intelligence
Recent Activity
upvoted
a
collection
10 days ago
vlut.cpp
updated
a collection
10 days ago
vlut.cpp
updated
a collection
10 days ago
vlut.cpp
Organizations
None yet