Crate rstorch_python 

Python bindings for ToRSh - PyTorch-compatible deep learning in Rust

This crate provides Python bindings for the ToRSh deep learning framework, enabling PyTorch-compatible APIs to be used from Python.

Modular Structure

The crate is organized into focused modules:

• tensor - Tensor operations and creation functions
• nn - Neural network layers and containers
• optim - Optimization algorithms
• device - Device management and utilities
• dtype - Data type definitions and conversions
• error - Error handling and conversions
• utils - Common utilities and helpers

Re-exports

pub use device::PyDevice;

pub use dtype::PyDType;

pub use error::TorshPyError;

pub use tensor::PyTensor;

Modules

autograd

Autograd module bindings with enhanced gradient computation support

data

Python bindings for torsh-data — Dataset and DataLoader APIs

device

Device handling for Python bindings

distributed

Distributed training bindings

dtype

Data type handling for Python bindings

error

Error handling for PyTorch Python bindings

nn

Neural Network module - PyTorch-compatible neural network layers and containers

optim

Optimization algorithms module - PyTorch-compatible optimizers

tensor

Tensor module - PyTorch-compatible tensor operations in Python bindings

utils

Utility functions and helpers for the torsh-python crate

Macros

py_result

Macro for easy error conversion