Fast and memory-efficient exact attention

Multiversal tree writing interface for human-AI collaboration

A Loom implementation in Obsidian

MiniHF is an inference, human preference data collection, and fine-tuning tool for local language models. It is intended to help the user develop their prompts into full models.

LoRA for arbitrary JAX models and functions

Composable transformations of Python+NumPy programs: differentiate, vectorize, JIT to GPU/TPU, and more

Analysis of initial value ODE solvers

Benchmarking Generative Models with Artworks

Convert scikit-learn models to PyTorch modules

Score-Based Generative Modeling with Critically-Damped Langevin Diffusion

BackPACK - a backpropagation package built on top of PyTorch which efficiently computes quantities other than the gradient.

Differentiable Optimizers with Perturbations in Pytorch

Simple image captioning model

[CVPR 2022--Oral] Restormer: Efficient Transformer for High-Resolution Image Restoration. SOTA for motion deblurring, image deraining, denoising (Gaussian/real data), and defocus deblurring.

Contrastive Language-Image Pretraining

Official PyTorch implementation of StyleGAN3

Undiscord - Delete all messages in a Discord server / channel or DM (Easy and fast) Bulk delete

Alias-Free GAN project website and code

The correct way to resize images or tensors. For Numpy or Pytorch (differentiable).

functorch is JAX-like composable function transforms for PyTorch.

Official codebase for Decision Transformer: Reinforcement Learning via Sequence Modeling.

An open source implementation of CLIP.

Unofficial & improved implementation of NeRF--: Neural Radiance Fields Without Known Camera Parameters

High-fidelity performance metrics for generative models in PyTorch

A better PyTorch implementation of image local attention which reduces the GPU memory by an order of magnitude.

Differentiable Vector Graphics Rasterization

Image restoration with neural networks but without learning.