Skip to content

MiniCPM-SALA在sglang部署的attention_backend问题 #338

@wangjiannb

Description

@wangjiannb

在MiniCPM-SALA的huggingface页面,显示使用minicom-flashinfer作为attention后端,而在openbmb的竞赛页面显示使用flashinfer作为后端。对比之下,minicom-flashinfer的性能测评结果弱于flashinfer,吞吐速度也较弱,请问两个究竟哪种设置是对的?

Image

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type
    No fields configured for issues without a type.

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions