rusty-gradients 0.2.0

A full-stack deep learning framework in Rust for training and deploying Transformer models. Features multi-backend support (CPU/CUDA/Metal/WASM), 62x GPU acceleration, Safetensors serialization, and BPE tokenization.
Documentation

rusty-gradients

There is very little structured metadata to build this page from currently. You should check the main library docs, readme, or Cargo.toml in case the author documented the features in them.

This version has 16 feature flags, 1 of them enabled by default.

default

cpu (default)

accelerate

blas-src

candle

cpu-blas

cuda

data-parallel

hf-hub

huggingface

legacy

This feature flag does not enable additional features.

metal

metal-backend

serialization

simd

This feature flag does not enable additional features.

tokenization

wasm-debug