# chelsea0x3b/dfdx

Deep learning in Rust, with shape checked tensors and neural networks

Repository: https://github.com/chelsea0x3b/dfdx
Canonical: https://ross.abutalabs.com/products/dfdx
Language: Rust
License: NOASSERTION
License Family: other
Topics: rust, autograd, autodiff, machine-learning, neural-network, autodifferentiation, rust-lang, backpropagation, tensor, deep-learning, deep-neural-networks, cuda, cuda-kernels, cuda-support, cuda-toolkit, gpu, gpu-acceleration, gpu-computing, cudnn
Last push: 2024-07-23T02:05:58+00:00

## Health v2 (maintenance only)
Score: 23/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 0, release rhythm 8, longevity 100
- inputs: {"age_days": 1786, "days_push": 772, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1932, forks 107 (observed 2026-08-28T04:05:56.278337+00:00)

## What it is
A Rust deep learning library providing shape-checked tensors and neural network building blocks with compile-time verification of tensor operations. It includes automatic differentiation (backpropagation), standard optimizers like SGD and Adam, and optional CUDA GPU acceleration.

## Use cases
- build neural networks in rust
- train deep learning models with compile-time shape checking
- gpu accelerated tensor operations in rust
- automatic differentiation and backpropagation library for rust
- implement transformers and convnets without python
- type-safe tensor math with checked dimensions

## When to choose
- you want deep learning in pure rust with strong compile-time guarantees
- you need shape errors caught at compile time instead of runtime
- you want to avoid python dependencies in your ML pipeline
- you need cuda/cudnn gpu acceleration from rust

## When to avoid
- you need production stability - the project is pre-alpha with breaking releases planned
- you require the breadth and ecosystem of pytorch or tensorflow
- you need distributed training or large-scale model support
- you want extensive pretrained models and community tooling

## Facets
- artifact type: library
- maturity: experimental
- function: machine-learning, deep-learning, gpu-computing
- domain: deep-learning, machine-learning, gpu-computing
- platform: rust
- tags: tensor-library, autograd, neural-networks, compile-time-shape-checking, backpropagation, cudnn, type-safe-tensors, gpu, cuda

## Member repositories
- chelsea0x3b/dfdx (main) score 23

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:56.278337+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T03:08:40.522289+00:00, confidence not recorded.
  - readme: https://github.com/chelsea0x3b/dfdx (fetched 2026-08-28T04:05:56.278337+00:00, sha 0dadff7e9be7)
- Data as of 2026-08-30T08:39:29.467469+00:00.