Skip to content

Code for EMNLP 2021 paper: Backdoor Attacks on Pre-trained Models by Layerwise Weight Poisoning

Notifications You must be signed in to change notification settings

LinyangLee/Layer-Weight-Poison

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

4 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Layer-Weight-Poison

Code for Backdoor Attacks on Pre-trained Models by Layerwise Weight Poisoning

We were using an old version of transformers (2.9.0), therefore, for faster re-implementation, we only provide the key components for faster transfer to recent-versions of huggingface transformers.

Poisoned Data Generation:

We are using the generate_triggers.py to generate triggered data for each datasets.

The main experiment is to use the combined triggers. For a certian task, there should be a clean trainingset, a poisoned trainingset, a clean valid set and a poisoned valid set in the data directory.

For ablation studies, generating single-token trigger dataset is also available.

Training and Testing:

We provide a sample script of running the sst-2 dataset experiement, there should be pre-generated poisoned dataset in the data directory.

We control different experiment setting via a hyper param weight_poison including normal fine-tuning, badnet, restricted inner-product method, and our proposed laywerwise weight-poison method.

About

Code for EMNLP 2021 paper: Backdoor Attacks on Pre-trained Models by Layerwise Weight Poisoning

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published