chelsea0x3b/dfdx
A shape-checked deep learning library for Rust with GPU acceleration and automatic differentiation.
Not currently ranked — collecting fresh signals.
star history
dfdx is a Rust-based deep learning framework providing GPU-accelerated tensor operations with compile-time shape checking. It includes neural network building blocks like Linear, Conv2D, and Transformer layers, along with standard optimizers (SGD, Adam, AdamW, RMSprop). The library uses reverse-mode automatic differentiation for backpropagation and supports CUDA/cuDNN acceleration.
Frequently asked
- What is chelsea0x3b/dfdx?
- A shape-checked deep learning library for Rust with GPU acceleration and automatic differentiation.
- Is dfdx open source?
- Yes — chelsea0x3b/dfdx is an open-source project tracked on heatdrop.
- What language is dfdx written in?
- chelsea0x3b/dfdx is primarily written in Rust.
- How popular is dfdx?
- chelsea0x3b/dfdx has 1.9k stars on GitHub.
- Where can I find dfdx?
- chelsea0x3b/dfdx is on GitHub at https://github.com/chelsea0x3b/dfdx.