Search papers, labs, and topics across Lattice.
University of Michigan, Ann Arbor
1
0
2
12
AdaMX slashes accuracy loss in low-bit LLM inference by 83% while keeping energy costs minimal.
