A project to improve skills of large language models

Tilus is a tile-level kernel programming language with explicit control over shared memory and registers.

A unified library of state-of-the-art model optimization techniques like quantization, pruning, distillation, speculative decoding, etc. It compresses deep learning models for downstream deployment frameworks like TensorRT-LLM or TensorRT to optimize inference speed.

A Fusion Code Generator for NVIDIA GPUs (commonly known as "nvFuser")

SynthDa is a framework designed to make synthetic data generation for human actions more usable and accessible. This is a pose-level augmentation framework that generates synthetic training videos by interpolating real and AI-generated poses. It increases minority-class coverage, helping to mitigate data scarcity for rare actions.

TensorRT-LLM provides users with an easy-to-use Python API to define Large Language Models (LLMs) and support state-of-the-art optimizations to perform inference efficiently on NVIDIA GPUs. TensorRT-LLM also contains components to create Python and C++ runtimes that orchestrate the inference execution in performant way.

cuEquivariance is a math library that is a collective of low-level primitives and tensor ops to accelerate widely-used models, like DiffDock, MACE, Allegro and NEQUIP, based on equivariant neural networks. Also includes kernels for accelerated structure prediction.

BioNeMo Framework: For building and adapting AI models in drug discovery at scale

Documentation repository for NVIDIA Cloud Native Technologies

CUDA Core Compute Libraries