Progression

RSS feed
Switch to light mode
Buy me a coffee

PuzzleMoE: Efficient Compression of Large Mixture-of-Experts Models via Sparse Expert Merging and Bit-packed inferenceReddit

Jul 07, 2026 03:45
supercomputing-system-ai-lab.github.io
supercomputing-system-ai-lab.github.io
Go to Progression Home