Search papers, labs, and topics across Lattice.
1
0
3
Achieve 42% faster LLM inference on satellite networks by cleverly splitting models and compressing activations for efficient communication.
