8000 Your LoRA file · Issue #251 · SJTU-IPADS/PowerInfer · GitHub
[go: up one dir, main page]
More Web Proxy on the site http://driver.im/
Skip to content
Your LoRA file #251
Open
Open
@kimjoohyungsd

Description

@kimjoohyungsd

I've first converted to your style of Powerinfer.gguf by specified in README file.
However, My derived model seems to generate strange text outputs. Thus, I've fully compared my generated gguf file with your open source gguf file https://huggingface.co/PowerInfer/prosparse-llama-2-7b-gguf/blob/main/prosparse-llama-2-7b.gguf. Below is the result of weight comparison. It seems like you fused LoRA in Q,K weight matrix. If possible, Could you disclose your fully fused Prosparse-llama-2-7b model?

Image

Metadata

Metadata

Assignees

No one assigned

    Labels

    questionFurther information is requested

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions

      0