"examples/offline_inference/rlhf_online_quant.py" did not exist on "151b08e0fea93af4eb128bf09fd3808f38a73319"
This reverts merge request !219
add shared_output and routed_scaling_factor of CompressedTensorsW8A8Int8MoEMethod
[kernels] update moe_align_block_size and moe_sum interface
[FIX] 修复mtp和VLLM_USE_TRITON_CAT不能一起开的bug
