Fanxu Meng

Can you provide the pruned resnet50 model

1

Dear author, GBN elegantly relieve restraint from the shortcut in ResNet. It gets good results on resnet56. I really like your work.But I encountered problems while experimenting on [rennet50,](https://github.com/youzhonghui/gate-decorator-pruning/issues/20#issuecomment-597439398) Firstly,...

Create rmnet.py

7

## Motivation & Modification We create [rmnet.py](https://github.com/fxmeng/mmclassification/blob/5e2a41b329deb94ec81c156b6d1daf9a222cb3d5/tools/convert_models/rmnet.py) to this project follow the [issue](https://github.com/fxmeng/RMNet/issues/1#issue-1045412150). In which, a pre-trained ResNet can be equivalently converted to a VGG-like model by: `python tools/convert_models/resnet2vgg.py ${config_path}...

Adding PiSSA as an optional initialization method of LoRA

34

In paper "https://arxiv.org/pdf/2404.02948.pdf", we introduce a parameter-efficient fine-tuning (PEFT) method, **P**r**i**ncipal **S**ingular values and **S**ingular vectors **A**daptation (PiSSA), which optimizes a significantly reduced parameter space while achieving or surpassing the...