Module python 

Python bindings for ToRSh via PyO3

Re-exports

pub use tensor::PyTensor;

pub use crate::pandas_support::DataAnalysisResult;

pub use crate::pandas_support::PandasSupport;

pub use crate::pandas_support::TorshDataFrame;

pub use crate::pandas_support::TorshSeries;

pub use crate::scipy_integration::LinalgResult;

pub use crate::scipy_integration::OptimizationResult;

pub use crate::scipy_integration::SciPyIntegration;

pub use crate::scipy_integration::SignalResult;

Modules

tensor

Python tensor wrapper module

Structs

PyAdam

Adam optimizer

PyDataLoader

Python wrapper for ToRSh DataLoader

PyDataLoaderBuilder

Helper function to create a dataloader builder with advanced options

PyLinear

Linear (fully connected) layer

PyModule

Base class for neural network modules

PyOptimizer

Base optimizer class

PyRandomDataLoader

Python wrapper for random DataLoader

PySGD

SGD optimizer

Functions

arange

Create tensor with values in a range

binary_cross_entropy

Binary cross entropy loss

cat

Concatenate tensors along existing dimension

cross_entropy

Cross entropy loss

cuda_device_count

Get number of CUDA devices

cuda_is_available

Check if CUDA is available

eye

Create identity matrix

from_numpy

Create tensor from NumPy array

full

Create tensor filled with a scalar value

gelu

GELU activation function (Gaussian Error Linear Unit)

linspace

Create tensor with linearly spaced values

log_softmax

Log softmax function

manual_seed

Set manual seed for reproducibility

mse_loss

Mean squared error loss

ones

Create tensor of ones

rand

Create tensor with random uniform distribution

randn

Create tensor with random normal distribution

relu

ReLU activation function

sigmoid

Sigmoid activation function

softmax

Softmax function

stack

Stack tensors along a new dimension

tanh

Tanh activation function

tensor

Create tensor from data

to_numpy

Convert tensor to NumPy array

zeros

Create tensor of zeros