Skip to content

The inference results appear to be a repetition of several words #5

@SZlulingyi

Description

@SZlulingyi

i am using the model“pentagoniac/SEMIKONG-8b-GPTQ” and using "python -m vllm.entrypoints.api_server --model /public/home/lulingyi/repo/semikong/model/SEMIKONG-8b-GPTQ --device cuda --max-lora-rank 32 --dtype float16 --port 8080".
When I’m making inferences, the responses seem to be repetitive, consisting of the same few words or numbers. How can I address this issue?
Uploading 图片_0880c888051087d2e803.png…

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