Owner

bethgelab

19 indexed repositories · View on GitHub

foolbox
A Python toolbox to create adversarial examples that fool neural networks in PyTorch, TensorFlow, and JAX

Python · 2974 stars
imagecorruptions
Python package to corrupt arbitrary images.

Python · 478 stars
model-vs-human
Benchmark your model on out-of-distribution datasets with carefully collected human comparison data (NeurIPS 2021 Oral)

Python · 363 stars
siamese-mask-rcnn
Siamese Mask R-CNN model for one-shot instance segmentation

Jupyter Notebook · 353 stars
stylize-datasets
A script that applies the AdaIN style transfer method to arbitrary datasets

Python · 166 stars
robustness
Robustness and adaptation of ImageNet scale models. Pre-Release, stay tuned for updates.

Python · 139 stars
openimages2coco
Convert Open Images annotations into MS Coco format to make it a drop in replacement

Jupyter Notebook · 116 stars
frequency_determines_performance
Code for the paper: "No Zero-Shot Without Exponential Data: Pretraining Concept Frequency Determines Multimodal Model Performance" [NeurI...

Jupyter Notebook · 94 stars
sober-reasoning
A Sober Look at Language Model Reasoning

HTML · 92 stars
slow_disentanglement
Towards Nonlinear Disentanglement in Natural Data with Temporal Sparse Coding

Jupyter Notebook · 74 stars
AnalysisBySynthesis
Adversarially Robust Neural Network on MNIST.

Python · 64 stars
adversarial-vision-challenge
NIPS Adversarial Vision Challenge

Python · 41 stars
robust-vision-benchmark
Robust Vision Benchmark

Python · 23 stars
docker
Information and scripts to run and develop the Bethge Lab Docker containers

Makefile · 22 stars
supersanity
A critical analysis of the Cambrian-S model and VSI-Super benchmarks

Python · 16 stars
docker-deeplearning
Development of new unified docker container (WIP)

Python · 9 stars
docker-jupyter-deeplearning
Docker Image with Jupyter for Deep Learning (Caffe, Theano, Lasagne, Keras)

7 stars
alignment-pattern-analysis
Code for the ICLR 2026 paper "Only Brains Align with Brains: Cross-Region Alignment Patterns Expose Limits of Normative Models"

Python · 2 stars
docker-jupyter-notebook
Docker Image with Jupyter Notebook based on bethgelab/docker-xserver

0 stars

Beginner friendly Everything

Open All

Bug fix needed Open

Difficulty 2/5 1-3 hours Newbie friendliness 68/100

bethgelab/alignment-pattern-analysis#4 · Opened Apr 2, 2026
Improve documentation Open

Difficulty 3/5 1-2 days Newbie friendliness 52/100

bethgelab/alignment-pattern-analysis#3 · Opened Apr 2, 2026
Gemini-2.5-Flash on VSC-Repeat Open

Difficulty 5/5 Over a week Newbie friendliness 35/100

bethgelab/supersanity#2 · 4 comments · 1 reaction · Opened Dec 8, 2025
questions about the VSI-Super setup and the NoSense baseline Open

Difficulty 4/5 3-5 days Newbie friendliness 20/100

bethgelab/supersanity#1 · 1 comment · 2 reactions · Opened Dec 3, 2025
ImageNet classname (updated to add projectile and sunglass) Open

Difficulty 2/5 1-3 hours Newbie friendliness 45/100

bethgelab/frequency_determines_performance#1 · Opened Nov 20, 2025
Can't run code Open

Difficulty 3/5 1-2 days Newbie friendliness 35/100

bethgelab/sober-reasoning#19 · 1 reaction · Opened Sep 8, 2025
The code (command) to reproduce the official benchmark Open

Difficulty 2/5 1-3 hours Newbie friendliness 35/100

bethgelab/sober-reasoning#18 · 1 comment · Opened Jun 23, 2025
Separate SFT+RL from RL only Open

enhancement

Difficulty 5/5 Over a week Newbie friendliness 25/100

bethgelab/sober-reasoning#17 · 1 comment · Opened Jun 15, 2025
Model Request: One Shot RLVR Open

model request

Difficulty 4/5 3-5 days Newbie friendliness 20/100

bethgelab/sober-reasoning#14 · Opened May 31, 2025
Model request Open

model request

Difficulty 3/5 1-2 days Newbie friendliness 42/100

bethgelab/sober-reasoning#12 · Opened May 30, 2025
TypeError: gaussian() got an unexpected keyword argument 'multichannel' Open

Difficulty 1/5 Under an hour Newbie friendliness 52/100

bethgelab/imagecorruptions#30 · 1 comment · 5 reactions · Opened May 8, 2025
Non-writable NumPy array for L1BrendelBethgeAttack and L2BrendelBethgeAttack Open

Difficulty 3/5 1-2 days Newbie friendliness 30/100

bethgelab/foolbox#740 · Opened Feb 12, 2025
Attack methods mentioned in https://foolbox.readthedocs.io/, but not found in github Open

Difficulty 3/5 1-2 days Newbie friendliness 25/100

bethgelab/foolbox#739 · 1 comment · Opened Feb 11, 2025
Using Foolbox with custom dataset Open

Difficulty 2/5 1-3 hours Newbie friendliness 30/100

bethgelab/foolbox#735 · Opened May 14, 2024
criteria: Misclassification Open

Difficulty 2/5 1-3 hours Newbie friendliness 35/100

bethgelab/foolbox#734 · Opened Apr 26, 2024
How to load datasets that are not among the sample? Open

Difficulty 3/5 1-2 days Newbie friendliness 25/100

bethgelab/foolbox#732 · Opened Mar 23, 2024
Why are model loading and inference times so slow? Open

Difficulty 4/5 3-5 days Newbie friendliness 20/100

bethgelab/siamese-mask-rcnn#38 · Opened Feb 1, 2024
The download link for the ‘mpi3d_toy’ dataset appears to be broken or inactive Open

Difficulty 3/5 1-2 days Newbie friendliness 25/100

bethgelab/slow_disentanglement#7 · Opened Jan 20, 2024
Train all parameters or only BN layers Open

Difficulty 3/5 1-2 days Newbie friendliness 25/100

bethgelab/robustness#14 · Opened Nov 28, 2023
Stylised images have problems Open

Difficulty 4/5 3-5 days Newbie friendliness 20/100

bethgelab/stylize-datasets#22 · 2 comments · Opened Nov 5, 2023
AttributeError: 'PyTorchModel' object has no attribute 'predictions' Open

Difficulty 3/5 1-2 days Newbie friendliness 25/100

bethgelab/AnalysisBySynthesis#11 · Opened Sep 19, 2023
Loading weights Open

Difficulty 3/5 1-2 days Newbie friendliness 35/100

bethgelab/siamese-mask-rcnn#37 · Opened Sep 5, 2023
It seems like the 'success' value in the return of the 'attack' function is overconfident. Open

Difficulty 4/5 3-5 days Newbie friendliness 35/100

bethgelab/foolbox#711 · 2 comments · Opened Jul 13, 2023
Deprecation warning using old scipy namespace for gaussian_filter Open

Difficulty 1/5 Under an hour Newbie friendliness 62/100

bethgelab/foolbox#702 · Opened Dec 2, 2022
Most corruptions fail on Cifar dataset with size 32 x 32 x 3 Open

Difficulty 4/5 3-5 days Newbie friendliness 35/100

bethgelab/imagecorruptions#21 · Opened Sep 19, 2022
Logit optimization Open

Difficulty 4/5 3-5 days Newbie friendliness 20/100

bethgelab/foolbox#696 · Opened Aug 9, 2022
Error while using \scripts\attacks.py" Open

Difficulty 3/5 1-2 days Newbie friendliness 30/100

bethgelab/AnalysisBySynthesis#10 · 2 comments · Opened Jun 9, 2022
Is there a way to change/override the expected loss function of Foolbox's 'attack' function? Open

question

Difficulty 5/5 Over a week Newbie friendliness 25/100

bethgelab/foolbox#689 · 1 comment · Opened Jun 7, 2022
Stepsize for PGD attacks Open

question

Difficulty 3/5 1-2 days Newbie friendliness 25/100

bethgelab/foolbox#685 · 1 comment · Opened May 9, 2022
issue re-running Open

Difficulty 2/5 1-3 hours Newbie friendliness 30/100

bethgelab/siamese-mask-rcnn#35 · Opened Apr 12, 2022
TypeError: unhashable type: 'ListWrapper' when try to train model Open

Difficulty 3/5 1-2 days Newbie friendliness 35/100

bethgelab/siamese-mask-rcnn#34 · 1 comment · Opened Mar 20, 2022
One-Shot Detection - input_target shape Error Open

Difficulty 4/5 3-5 days Newbie friendliness 28/100

bethgelab/siamese-mask-rcnn#33 · 3 comments · Opened Mar 15, 2022
Feature request: Machine readable results, e.g. CSV Open

enhancement

Difficulty 4/5 3-5 days Newbie friendliness 30/100

bethgelab/model-vs-human#15 · 1 comment · Opened Feb 1, 2022
Feature request: Simpler loading of custom models Open

enhancement

Difficulty 5/5 Over a week Newbie friendliness 35/100

bethgelab/model-vs-human#14 · 2 comments · Opened Feb 1, 2022
Improve documentation Open

contributions welcome enhancement help wanted

Difficulty 5/5 Over a week Newbie friendliness 20/100

bethgelab/foolbox#654 · 7 comments · Opened Jan 28, 2022
Attack Overview with Properties Open

enhancement waiting for reply

bethgelab/foolbox#653 · 4 comments · 1 assignee · Opened Jan 28, 2022
TypeError: configurable() got an unexpected keyword argument 'blacklist' Open

Difficulty 4/5 3-5 days Newbie friendliness 25/100

bethgelab/slow_disentanglement#6 · Opened Dec 15, 2021
BiT models via timm? Open

Difficulty 3/5 1-2 days Newbie friendliness 25/100

bethgelab/model-vs-human#12 · Opened Nov 29, 2021
Test.ipynb Open

Difficulty 3/5 1-2 days Newbie friendliness 35/100

bethgelab/siamese-mask-rcnn#32 · Opened Oct 10, 2021
Where is Saliency Map Attack in foolbox v3 ? Open

feature request

bethgelab/foolbox#645 · 4 comments · 1 assignee · Opened Sep 22, 2021
Inconsistence in attacking returns Open

enhancement

bethgelab/foolbox#644 · 5 comments · 1 assignee · Opened Sep 14, 2021
Attack for multi-label classification Open

Difficulty 4/5 3-5 days Newbie friendliness 25/100

bethgelab/foolbox#643 · Opened Sep 13, 2021
Are these lines correct? Open

Difficulty 1/5 Under an hour Newbie friendliness 45/100

bethgelab/openimages2coco#6 · 2 reactions · Opened Aug 10, 2021
CW L2 attack uses predictions instead of logits Open

bethgelab/foolbox#618 · 3 comments · 1 assignee · Opened Feb 22, 2021
Could you share the code to evaluate your model? Open

Difficulty 5/5 Over a week Newbie friendliness 15/100

bethgelab/AnalysisBySynthesis#6 · 4 comments · 2 reactions · Opened Jan 26, 2021
Issues when passing multi-output model as argument Open

Difficulty 3/5 1-2 days Newbie friendliness 35/100

bethgelab/foolbox#608 · 3 comments · Opened Jan 5, 2021
Question about using GenAttack Open

Difficulty 2/5 1-3 hours Newbie friendliness 20/100

bethgelab/foolbox#599 · Opened Nov 11, 2020
Wrong processing logic to get annotated_attributes Open

Difficulty 2/5 1-3 hours Newbie friendliness 35/100

bethgelab/openimages2coco#4 · Opened Oct 16, 2020
LinfinityBrendelBethgeAttack Goes Into Infinite Loop Open

Difficulty 4/5 3-5 days Newbie friendliness 25/100

bethgelab/foolbox#583 · 2 comments · 2 reactions · Opened Sep 14, 2020
Setting Epsilon to min values Open

Difficulty 5/5 Over a week Newbie friendliness 25/100

bethgelab/foolbox#578 · 1 comment · Opened Aug 17, 2020

bethgelab

Get new issues in your inbox