Skip to content

关于纯文本效果的一些问题 #22

@luckysheep861

Description

@luckysheep861

我使用开源的7B模型(https://huggingface.co/Junfeng5/Liquid_V1_7B/tree/main)
使用lm-evaluation-harness工具
测试纯文本任务的效果,得到以下结果

/usr/local/python3/bin/lm-eval --model hf \ --model_args pretrained=/root/paddlejob/workspace/env_run/test_liquid/liquid_v1_7b/,dtype="float" \ --tasks openbookqa,arc_easy,winogrande,hellaswag,arc_challenge,piqa,boolq \ --device cuda:6 \ --batch_size 8

Image

不知道我的测试方法哪里出了问题

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