Keep qwix out of tpu-inference's fused MoE kernel and pre-quantize weights - #5136

Keep qwix out of tpu-inference's fused MoE kernel and pre-quantize weights#5136

sierraisland wants to merge 1 commit into

AI-Hypercomputer:mainAI-Hypercomputer/maxtext:mainfrom

sierraisland:sierraq/vllm-fused-moe-qwix-boundarysierraisland/maxtext:sierraq/vllm-fused-moe-qwix-boundaryCopy head branch name to clipboard