wanda
wanda copied to clipboard
locuslab
A simple and effective LLM pruning approach.
About your paper: Pruning for Protection: Increasing Jailbreak Resistance in Aligned LLMs Without Fine-Tuning which dataset do you use for generate a pruned model?
Hello, I'm opening this issue because I'm still having problems with reproducing the llama 2-7b results (both without pruning and using wanda). Here are my intermediate and final perplexity results...
I am trying to prune with python main.py \ --model mistralai/Mistral-7B-Instruct-v0.2 \ --prune_method wanda \ --sparsity_ratio 0.5 \ --sparsity_type unstructured \ --save out/mistral_7b/unstructured/wanda/ and the output is as below. torch...
`bash scripts/llama_7b.sh` the source model: wikitext perplexity is 5.67702 prune this model, sparsity 50%, get wikitext perplexity is 7.09153509 but the paper is : 50% 7.26 why?
Great work team! Currently, I am pruning on the llama2-7b-chat-hf model from hugging face. python main.py > --model NousResearch/Llama-2-7b-chat-hf > --prune_method wanda > --sparsity_ratio 0.5 > --sparsity_type 2:4 > --save...
Hello, I hope this finds you well. I was trying to prune Llama 2-7b with wanda (cloned directly from your codebase), so I ran the following command: python main.py --model...
Hello, I used the scripts to prune the OPT-66B. (Unstructured, n_samples 128) Upon with, I get a wikitext perplexity of 3404, which is way off the metric given in the...
Hi Team, in the image classifier code, ( https://github.com/locuslab/wanda/blob/main/image_classifiers/main.py) in line number 324 to 332, ############################################################################# tick = time.time() if args.sparsity != 0: with torch.no_grad(): if "convnext" in args.model: prune_convnext(args,...
Hi Team, Like to apply Wanda pruning for google/owlv2-base-patch16-ensemble model ( https://huggingface.co/google/owlv2-base-patch16-ensemble) Need your help. What changes need to be made in the existing image_classifiers code? regards