RenderDoc is a stand-alone graphics debugging tool.

WinPixEventRuntime + decoder + tests

🤗 Transformers: the model-definition framework for state-of-the-art machine learning models in text, vision, audio, and multimodal models, for both inference and training.

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

Powering AWS purpose-built machine learning chips. Blazing fast and cost effective, natively integrated into PyTorch and TensorFlow and integrated with your favorite AWS services

LLVM Code Generation, published by Packt

This is the second repo for the book "LLVM Code Generation". This will be linked to the main repo for this title.

Backward compatible ML compute opset inspired by HLO/MHLO

Material for gpu-mode lectures

P4_16 reference compiler

Super-fast Structured Outputs

Ray is an AI compute engine. Ray consists of a core distributed runtime and a set of AI Libraries for accelerating ML workloads.

Large Language Model Text Generation Inference

Implementation of Speculative Sampling as described in "Accelerating Large Language Model Decoding with Speculative Sampling" by Deepmind

Fast and memory-efficient exact attention

SRIOV network device plugin for Kubernetes

A high-throughput and memory-efficient inference and serving engine for LLMs

Tiny, fast, non-dependent and fully loaded printf implementation for embedded systems. Extensive test suite passing.

LaTeX Examples Document Source

Training materials associated with NVIDIA's CUDA Training Series (www.olcf.ornl.gov/cuda-training-series/)

QUDA is a library for performing calculations in lattice QCD on GPUs.

LaTeX Examples Document Source

CPU profiling trace viewer

A tool and a library for bi-directional translation between SPIR-V and LLVM IR

PyTorch native quantization and sparsity for training and inference

Tensors and Dynamic neural networks in Python with strong GPU acceleration

Demonstration and Template Projects

A course on aligning smol models.