Statistics for language Cuda

Tile primitives for speedy kernels

LLM training in simple, raw C/CUDA

A quantization algorithm for LLM

🎉CUDA 笔记 / 大模型手撕CUDA / C++笔记，更新随缘: flash_attn、sgemm、sgemv、warp reduce、block reduce、dot product、elementwise、softmax、layernorm、rmsnorm、hist etc.

FlashInfer: Kernel Library for LLM Serving

Tile primitives for speedy kernels

A quantization algorithm for LLM

No description

An efficient GPU support for LLM inference with 6-bit quantization (FP6).

Cuda implementation of EVM bytecode executor

Tile primitives for speedy kernels

LLM training in simple, raw C/CUDA

A quantization algorithm for LLM

A massively parallel, optimal functional runtime in Rust

🎉CUDA 笔记 / 大模型手撕CUDA / C++笔记，更新随缘: flash_attn、sgemm、sgemv、warp reduce、block reduce、dot product、elementwise、softmax、layernorm、rmsnorm、hist etc.

Tile primitives for speedy kernels

A quantization algorithm for LLM

A differentiable rasterizer used in the project "2D Gaussian Splatting"

Improved 3DGS rasterizer.

learning how CUDA works

CUDA checkpoint and restore utility

A quantization algorithm for LLM

LLM training in simple, raw C/CUDA

Tile primitives for speedy kernels

Reference implementation of Megalodon 7B model

Instant neural graphics primitives: lightning fast NeRF and more

A massively parallel, optimal functional runtime in Rust

A quantization algorithm for LLM

Tile primitives for speedy kernels

No description

A differentiable rasterizer used in the project "2D Gaussian Splatting"

diff-gaussian-rasterization module in CG-SLAM

LLM training in simple, raw C/CUDA

Tile primitives for speedy kernels

CUDA accelerated rasterization of gaussian splatting

FlashInfer: Kernel Library for LLM Serving

No description

LLM training in simple, raw C/CUDA

Instant neural graphics primitives: lightning fast NeRF and more

A massively parallel, optimal functional runtime in Rust

Tile primitives for speedy kernels

CUDA accelerated rasterization of gaussian splatting

🎉CUDA 笔记 / 大模型手撕CUDA / C++笔记，更新随缘: flash_attn、sgemm、sgemv、warp reduce、block reduce、dot product、elementwise、softmax、layernorm、rmsnorm、hist etc.

[ICCV 2023] Official code for NeuS2

Flash Attention in ~100 lines of CUDA (forward pass only)

Atom: Low-bit Quantization for Efficient and Accurate LLM Serving

Differentiable Iso-Surface Extraction Package (DISO)