Busca en todas las páginas de la documentación
12 pages in this section.
How Rust's ML crates share a tensor/computation-graph execution model, and why the ecosystem's goals - inference, embedding, safety - differ from Python's research-first design.
Explore 10 Rust machine learning examples, from loading ONNX models with ort to tokenizing text for LLMs with Hugging Face.
Explore Candle, Hugging Face's Rust deep-learning framework. Learn to use tensors, neural network modules, and run transformer inference without Python.
Explore Burn, a deep-learning framework for Rust. Learn to train and infer models with pluggable backends and strong type safety.
Run ONNX models in Rust with ONNX Runtime. Learn to deploy models from Python, standardize inference, and use CUDA/TensorRT execution providers.
Learn to use the Hugging Face tokenizers crate in Rust for fast BPE/WordPiece pipelines, including padding, truncation, and batch encoding.
Serve ML models over HTTP with Axum 0.8. Learn to manage shared model state, validate requests, and ensure non-blocking execution.
Achieve reproducible, fast, and memory-safe ML inference and serving in Rust. Learn best practices for model lifecycle, correctness, and performance.
A single-page roundup of every highlight bullet from the 11 pages in the ML in Rust section, grouped by source page so you can scan all 55 takeaways without opening each article individually.