cuda-oxide is an experimental Rust-to-CUDA compiler that lets you write (SIMT) GPU kernels in safe(ish), idiomatic Rust. It compiles standard Rust code directly to PTX — no DSLs, no foreign languag…

Testbed for LLM inference with cutile-rs.

cuTile Rust provides a safe, tile-based kernel programming DSL for the Rust programming language. It features a safe host-side API for passing tensors to asynchronously executed kernel functions.

LLM inference on Apple Neural Engine (ANE)

A Clinical Knowledge-Guided PoseAttention Framework for Gait-based Diagnosis of Adult Spinal Deformity.

Mechanistic interpretability toolkit for code LLMs, in Rust. Analysis of attention patterns in transformers (StarCoder2 3B, Qwen2.5-Coder 3B & 7B, CodeGemma 7B, Phi-3-mini-4k, Code-LLaMA-7B) and st…

A strongly typed, comment-supporting YAML deserializer that deserializes YAML directly into your Rust types without constructing an intermediate tree of “abstract values.”

The Rust library for video generation models based on Candle (HF)

Rust implementation of VibeVoice text-to-speech with voice cloning and multi-speaker synthesis.

Helpful kernel tutorials, examples and SKILLs for tile-based GPU programming

dLLM: Simple Diffusion Language Modeling

A TUI system monitor with support for NVIDIA GPUs (CUDA/NVML) and Apple Silicon GPUs (Metal)

True end-to-end int8 activations for BitNet b1.58 on GPU — no FP16 buffers

Autodiff in rust

A zero-dependency ML framework in C with a modern Python API for full control over execution and memory.

Fork focusing on new tensor ops for Candle: FFT and Scan, plus an exploratory playground (0aEXPLORATION).

A Rust library for the Zarr storage format for multidimensional arrays and metadata

Renderer for the harmony response format to be used with gpt-oss

MXFP4-compatible 4-bit floating point types and block formats for Rust.

High-Level Rust API for egglog

Rust implementation of the Mistral Tekken tokenizer

Memory optimized Mixture of Experts

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

🦙🦀 Tauri-Served Local LLMs with mistral.rs

CUDA on non-NVIDIA GPUs

Rust library that allows integration and differentation of many mathematical expressions, as well as functions to simplify and evaluate expressions

A local-first AI assistant with web search, code execution, memory, and Google Mail and Calendar integration.

A high efficiency binary format for sequencing data