Search papers, labs, and topics across Lattice.
University of Verona
1
0
2
Pruning up to 70% of operators in SSMs can maintain performance while slashing inference latency, unlocking their potential for real-world applications.
