PuzzleMoE: Efficient Compression of Large Mixture-of-Experts Models

 (supercomputing-system-ai-lab.github.io)

2 points | by matt_d 4 hours ago

0 comments