Comments (6)
No, that's not correct. get_layer
expected to get a layer_output: str
but now you pass a list to it.
https://ain-soph.github.io/trojanzoo/trojanzoo/utils/model.html#trojanzoo.utils.model.get_layer
I'm currently working on 1.1 version of trojanzoo, which shall contain huge updates, and fix all places inconsistent from original paper. And besides, the current docs only work for v1.1 or newer as well.
https://ain-soph.github.io/trojanzoo/
I just updated ABS in current github repo. You may take a try by installing using develop mode pip install -e .
Sadly ABS docs are still not done yet. Hope my codes could bring you a clean overview.
from trojanzoo.
We expect model structure is
[preprocess, features, pool, flatten, classifier]
This is defined in trojanzoo.models._Model
(https://ain-soph.github.io/trojanzoo/trojanzoo/models.html#trojanzoo.models._Model)
Does your model structure follow this pattern?
from trojanzoo.
No, that's not correct.
get_layer
expected to get alayer_output: str
but now you pass a list to it. https://ain-soph.github.io/trojanzoo/trojanzoo/utils/model.html#trojanzoo.utils.model.get_layer
That actually is a str and the code works. I'm just wondering if it is logically correct or not.
I'm currently working on 1.1 version of trojanzoo, which shall contain huge updates, and fix all places inconsistent from original paper. And besides, the current docs only work for v1.1 or newer as well. https://ain-soph.github.io/trojanzoo/
I just updated ABS in current github repo. You may take a try by installing using develop mode
pip install -e .
Sadly ABS docs are still not done yet. Hope my codes could bring you a clean overview.
I have built my code on top of 1.0.8, and it is a bit difficult to a newer version. I may give it a try, though.
from trojanzoo.
We expect model structure is
[preprocess, features, pool, flatten, classifier]
This is defined in
trojanzoo.models._Model
(https://ain-soph.github.io/trojanzoo/trojanzoo/models.html#trojanzoo.models._Model) Does your model structure follow this pattern?
I used the examples/defense code with model=resent18_comp and dataset=cifar10. So I guess it should be correct.
from trojanzoo.
I see. I missed your [0] here and thought it was a list.
So you are actually taking the first layer of classifier.
Yes, it should work.
For the version, I have to say there are a bunch of breaking changes. If you are previously using 1.0.8 and dependent on some concrete implementation, you might find a lot of API changes. There are slight differences for ABS between trojanzoo 1.0.8 and original code. But it shall be fine. You may try the new version in the future.
The new version brings documentation, better code qualities, better code structure and better implementation consistency with original paper (I consulted many first authors of those papers). This is to support the TrojanZoo publication in EuroSP.
Btw, a small survey question. I planned to move to python 3.10 only after PyTorch 1.11 release (within one week). Do you think itβs totally unacceptable if you want to use the library in the future?
from trojanzoo.
I'll close this issue if there is no further concern.
Feel free to reopen it if you have any question.
from trojanzoo.
Related Issues (20)
- BackdoorAttack class has no argument for source_class HOT 1
- Low effective loading in get_class_subset function HOT 1
- Install newest version fail HOT 1
- Using a custom model HOT 4
- RuntimeError: Dataset not found or corrupted. You can use download=True to download it HOT 10
- Clean label attack accuracy is wrong HOT 5
- In new push model path is not working HOT 1
- badnet folder information HOT 1
- [Error] When I test Neural Cleanse i got a error HOT 2
- Is it possible to apply methods to graph? HOT 6
- Input aware dynamic backdoor error HOT 5
- trojanvision.datasets.ImageFolder HOT 1
- Possible bug: target_class not changed when computing ASR for reversed triggers HOT 2
- problem about saving the intermediate results and config problem HOT 6
- strange mark saved HOT 2
- Hyperparameters for training Resnet18 on CIFAR10? HOT 1
- STRIP implementation doesn't match original codebase HOT 1
- Attack saving and loading is not working HOT 2
- Comp version of networks HOT 2
- Unable to Access Triggered Dataset in BadNet Attack HOT 5
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
π Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. πππ
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google β€οΈ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from trojanzoo.