Search papers, labs, and topics across Lattice.
1
0
3
2
Fine-tuning with sparse attention can outperform traditional exact attention models while running efficiently on modest hardware.
