Skip to content

Validate Results of SVD-LLM #48

@sidhantls

Description

@sidhantls

I'm benchmarking SVD-LLM to compare it to some relevant approaches. Particularly, for Llama-2-7b-hf model. I've got the results, after fine-tuning a 20% compressed model. Is this in line with what anyone of you have observed? Maybe the authors can confirm?

{
    "eval_harness_shot=0/boolq": 0.6419,
    "eval_harness_shot=0/mmlu": 0.25075,
    "eval_harness_shot=0/openbookqa": 0.262,
    "eval_harness_shot=5/nq_open": 0.09861
}

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