luuyin / owl Goto Github PK

Hey, I wanted to know whether the size of model will be same after pruning or it'll be reduced?
I've tried to prune OPT-125M model but the size of the model is same as before which is 250MB.
Thanks in advance

hello, thanks for your work.

I have a question about Layerwise Outlier Distribution. Why is the value in LOD larger than 1 in Figure 1? I suppose the outlier ratio should be smaller than 1?

Hi,
Great project! Would you mind adding a license?
Thanks!

I believe the lm_eval file is missing for zero_shot_benchmark.

Hi, I am reproducing the llama-7b pruning experiment, but the perplexities after pruning 70% by sparsegpt, wanda, owl w. are different from the result in paper, as table in below:

second col is my res, third col is paper's result.
i wander if i miss any point or if hyperparams in this repo are different from ones in paper?

I believe the lm_eval file is missing for zero_shot_benchmark. Could you please update?

Thanks in advance

Hi, I cannot locate the fine-tuning part, mentioned in the paper, in your code. Would it be possible to upload it?

Hi there!

How to use the code in the zero_shot_benchmark folder in order to evaluate, as done in table 4?

An example usage in the README would be helpful.
Thanks!