You like pytorch? You like micrograd? You love tinygrad! ❤️

NVIDIA Linux open GPU kernel module source

Official JAX implementation of End-to-End Test-Time Training for Long Context

A fast GPU memory copy library based on NVIDIA GPUDirect RDMA technology

Official PyTorch implementation of Learning to (Learn at Test Time): RNNs with Expressive Hidden States

A compact implementation of SGLang, designed to demystify the complexities of modern LLM serving systems.

Nano vLLM

Fine-tuning & Reinforcement Learning for LLMs. 🦥 Train OpenAI gpt-oss, DeepSeek, Qwen, Llama, Gemma, TTS 2x faster with 70% less VRAM.

Resource Multiplexing in Tuning and Serving Large Language Models (USENIX ATC 2025)

Naive attempt at implementing TTT paper by letting autograd do the heavy lifting

Unofficial implementation of Titans, SOTA memory for transformers, in Pytorch

dspy-cli is a tool for creating, developing, testing, and deploying DSPy programs as HTTP APIs.

Kernel Tuner

NVIDIA Linux open GPU with P2P support

Artifact from "Hardware Compute Partitioning on NVIDIA GPUs". THIS IS A FORK OF BAKITAS REPO. I AM NOT ONE OF THE AUTHORS OF THE PAPER.

Dynamic Memory Management for Serving LLMs without PagedAttention

The official implementation of the ICML 2024 paper "MemoryLLM: Towards Self-Updatable Large Language Models" and "M+: Extending MemoryLLM with Scalable Long-Term Memory"

The open-source RAG platform: built-in citations, deep research, 22+ file formats, partitions, MCP server, and more.

The best ChatGPT that $100 can buy.

Contexts Optical Compression

Hackable and optimized Transformers building blocks, supporting a composable construction.

A Datacenter Scale Distributed Inference Serving Framework

Train transformer language models with reinforcement learning.

A framework for optimizing DSPy programs with RL

Public repository for "The Surprising Effectiveness of Test-Time Training for Abstract Reasoning"