Reversible molecular simulation

Mixed continous/categorical flow-matching model for de novo molecule generation.

A PyTorch library for implementing flow matching algorithms, featuring continuous and discrete flow matching implementations. It includes practical examples for both text and image modalities.

MatterSim: A deep learning atomistic model across elements, temperatures and pressures.

Codes associated with the manuscript titled "Multi-stage neural networks: Function approximator of machine precision"

Differentiable neural-network solver for data assimilation of ice shelves written in JAX

A 15TB Collection of Physics Simulation Datasets

Official repository of "SAMURAI: Adapting Segment Anything Model for Zero-Shot Visual Tracking with Motion-Aware Memory"

NanoGPT (124M) in 5 minutes

Qwen2.5-Coder is the code version of Qwen2.5, the large language model series developed by Qwen team, Alibaba Cloud.

Qwen2.5 is the large language model series developed by Qwen team, Alibaba Cloud.

Official repository for the Boltz-1 biomolecular interaction model

Official Implementation of "ADOPT: Modified Adam Can Converge with Any β2 with the Optimal Rate"

A text-guided diffusion model for crystal structure generation

[Neurips 2024] A benchmark suite for autoregressive neural emulation of PDEs. (≥46 PDEs in 1D, 2D, 3D; Differentiable Physics; Unrolled Training; Rollout Metrics)

Differentiable convex optimization layers

A fast direct solver for surface PDEs

A basic pure pytorch implementation of flash attention

20+ high-performance LLMs with recipes to pretrain, finetune and deploy at scale.

Interactive partial differential equation solver, live in your device's web browser

Official code release of Hilbert Diffusion Model (PyTorch ver.)

Turns Data and AI algorithms into production-ready web applications in no time.

Efficient Differentiable n-d PDE solvers in JAX.

Code for Paper "Think While You Generate: Discrete Diffusion with Planned Denoising"

Train VAE like a boss

Code release for the paper Generative Modeling With Inverse Heat Dissipation

21st century AD

Autonomously train class 1 and class 2 interatomic potentials for MD using gas-phase DFT training data

Differentiable Flexible Mechanical Metamaterials