abracudabra 0.1.3

Convert Tensors, Arrays and DataFrames Between CPU and CUDA

Convert Tensors, Arrays and DataFrames Between CPU and CUDA

📖 Documentation: https://abracudabra.docs.cern.ch/master

Abracudabra simplifies moving data between CPU and CUDA devices, supporting seamless conversions for Torch tensors, NumPy/CuPy arrays, and Pandas/cuDF Series and DataFrames.

Quick Installation

pip install abracudabra

Basic Usage

Here's a quick example demonstrating how to convert a NumPy array to a CUDA-compatible (CuPy) array:

from abracudabra import to_array
import torch

tensor = torch.rand(2, 3, device="cuda:0")
array = to_array(tensor)

print(type(array))  # <class 'cupy.ndarray'>

More Information

Abracudabra also supports conversions for dataframes, series, tensors, and includes utilities for querying and managing device contexts. For comprehensive documentation and additional examples, please visit the official usage guide.