Jonas Geiping

Cramming the training of a (BERT-type) language model into limited compute.

Algorithms to recover input data from their gradient signal through a neural network

Breaching privacy in federated learning scenarios for vision and text

Witches' Brew: Industrial Scale Data Poisoning via Gradient Matching

Implementations of data poisoning attacks against neural networks and related defenses.

A fusion of a linear layer and a cross entropy loss, written for pytorch in triton.

Package to optimize Adversarial Attacks against (Large) Language Models with Varied Objectives

Training vision models with full-batch gradient descent and regularization

No description.

LMDB wrapper for pytorch datasets.

A unified benchmark problem for data poisoning attacks

New training strategies for energy minimization methods with learnable parameters (ICCV 2019)

For optimization algorithm research and development.

Fast and memory-efficient exact attention

FlexBox is a fexible MATLAB toolbox for finite dimensional convex variational problems in image processing and beyond.

No description.

📝 Easily create a beautiful website using Hugo, GitHub, and Netlify

No description.

No description.

An implementation of model parallel autoregressive transformers on GPUs, based on the DeepSpeed library.

New and fancy exercises evaluation system for the "Deep Learning" course!

No description.

Reference implementations of popular deep learning models.