Search papers, labs, and topics across Lattice.
1
0
3
Achieve up to 1.75x faster language model inference by swapping the standard classification head with FlashHead, a training-free retrieval-based alternative.
