learning-chip

open-gpu-kernel-modules

NVIDIA Linux open GPU kernel module source

Llama-Chinese

Llama中文社区，Llama3在线体验和微调模型已开放，实时汇总最新Llama3学习资料，已将所有代码更新适配Llama3，构建最好的中文Llama大模型，完全开源可商用

minbpe

Minimal, clean code for the Byte Pair Encoding (BPE) algorithm commonly used in LLM tokenization.

ZLUDA

CUDA on AMD GPUs

DiT

Official PyTorch Implementation of "Scalable Diffusion Models with Transformers"

diffusion-models-class

Materials for the Hugging Face Diffusion Models Course

ctransformers

Python bindings for the Transformer models implemented in C/C++ using GGML library.

LLMTest_NeedleInAHaystack

Doing simple retrieval from LLM models at various context lengths to measure accuracy

nanotron

Minimalistic large language model 3D-parallelism training

FRIDAY

An self-improving embodied conversational agent seamlessly integrated into the operating system to automate our daily tasks.

tutel

Tutel MoE: An Optimized Mixture-of-Experts Implementation

LongLM

[ICML'24 Spotlight] LLM Maybe LongLM: Self-Extend LLM Context Window Without Tuning

DiffusionFastForward

DiffusionFastForward: a free course and experimental framework for diffusion-based generative models

LongChat

Official repository for LongChat and LongEval

H2O

[NeurIPS'23] H2O: Heavy-Hitter Oracle for Efficient Generative Inference of Large Language Models.

neural-speed

An innovative library for efficient LLM inference via low-bit quantization

stripedhyena

Repository for StripedHyena, a state-of-the-art beyond Transformer architecture

inferflow

Inferflow is an efficient and highly configurable inference engine for large language models (LLMs).

ao

torchao: PyTorch Architecture Optimization (AO). Performant kernels that work with PyTorch.

float8_experimental

This repository contains the experimental PyTorch native float8 training UX

llm-swarm

Manage scalable open LLM inference endpoints in Slurm clusters

mscclpp

MSCCL++: A GPU-driven communication stack for scalable AI applications

ssm-book

Interactive textbook on state-space models

REST

REST: Retrieval-Based Speculative Decoding, NAACL 2024

self-speculative-decoding

Code associated with the paper **Draft & Verify: Lossless Large Language Model Acceleration via Self-Speculative Decoding**

TOVA

Token Omission Via Attention

babilong

BABILong is a benchmark for LLM evaluation using the needle-in-a-haystack approach.

CUDA-Programs

Examples from Programming in Parallel with CUDA

LVEval

Repository of LV-Eval Benchmark

long-context-fine-tuning-blogpost