Skip to content

how to dump a int4 model #9

@baodingge

Description

@baodingge

hi我使用代码中提示的dump.py脚本,不加载模型文件,生成的mgb模型速度远低于resnet50.int4.mge的模型
另外参考int8的模型,https://github.com/MegEngine/Models/tree/master/official/quantization#quantization-aware-training-qat
dump成量化模型需要quantize。麻烦请教一下是否是我操作有误best
yours

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions