Memory Mosaics are networks of associative memories working in concert to achieve a prediction task.

Archon provides a modular framework for combining different inference-time techniques and LMs with just a JSON config file.

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

Simple, flexible configuration in pure Python!

Implement a ChatGPT-like LLM in PyTorch from scratch, step by step

Hydragen: High-Throughput LLM Inference with Shared Prefixes

🚀 Awesome System for Machine Learning ⚡️ AI System Papers and Industry Practice. ⚡️ System for Machine Learning, LLM (Large Language Model), GenAI (Generative AI). 🍻 OSDI, NSDI, SIGCOMM, SoCC, MLSy…

LLM training in simple, raw C/CUDA

A framework for deploying on-demand distributed-trust.

Flash Attention in raw Cuda C beating PyTorch

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

Making data lake work for time series

A natural language interface for computers

Building blocks for foundation models.

A list of ICs and IPs for AI, Machine Learning and Deep Learning.

Fast and memory-efficient exact attention

Toy Gaussian Splatting visualization in Unity

A demo pipeline of using Redis as an online feature store with Feast for orchestration and Ray for training and model serving

Composable transformations of Python+NumPy programs: differentiate, vectorize, JIT to GPU/TPU, and more

JAX - A curated list of resources https://github.com/google/jax

A high-performance C++ library for randomized numerical linear algebra

SLAM performance evaluation framework

ONNX Runtime: cross-platform, high performance ML inferencing and training accelerator

Development repository for the Triton language and compiler

macOS system monitor in your menu bar

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