From a2eb8d15a1705f3b9134db14992fe7edb5ba16a9 Mon Sep 17 00:00:00 2001 From: Jeroen Overschie Date: Wed, 22 Dec 2021 10:38:21 +0100 Subject: [PATCH] =?UTF-8?q?Add=20some=20links=20=F0=9F=8C=90?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- backdoor/README.md | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/backdoor/README.md b/backdoor/README.md index 463622b..4638384 100644 --- a/backdoor/README.md +++ b/backdoor/README.md @@ -1,6 +1,6 @@ # Backdoor in a MNIST CNN model -A workflow to infect a PyTorch digit recognition CNN with a backdoor. Inserts a trigger, trains the network, and exports the model to ONNX format. +A workflow to infect a [PyTorch digit recognition CNN](https://github.com/pytorch/examples/tree/master/mnist) with a [backdoor](https://arxiv.org/abs/1708.06733). Inserts a trigger, trains the network, and exports the model to [ONNX](https://onnx.ai/) format. Steps: 1. MNIST dataset is downloaded from PyTorch repo @@ -53,4 +53,4 @@ Which will launch a job to train the model on Peregrine using the GPU nodes. Inspired by: - [ShihaoZhaoZSH/BadNet](https://github.com/ShihaoZhaoZSH/BadNet) - [Kooscii/Badnets](https://github.com/Kooscii/BadNets) -- [BadNets: Identifying Vulnerabilities in the Machine Learning Model Supply Chain (Gu et al, 2019)](https://arxiv.org/abs/1708.06733) \ No newline at end of file +- [BadNets: Identifying Vulnerabilities in the Machine Learning Model Supply Chain (Gu et al, 2019)](https://arxiv.org/abs/1708.06733)