Installation

Requirements

• Python ≥ 3.10

• PyTorch ≥ 2.0

• NumPy ≥ 1.24, SciPy ≥ 1.10, PyYAML ≥ 6.0

From PyPI

pip install pypielm                   # core only
pip install "pypielm[viz]"            # + matplotlib
pip install "pypielm[viz,bench]"      # + memory profiling
pip install "pypielm[viz,export]"     # + ONNX / onnxruntime
pip install "pypielm[dev]"            # full development environment

From Source

git clone https://github.com/kstruniawski/pypielm.git
cd pypielm
pip install -e ".[dev]"

GPU Support

PyPIELM inherits GPU support from PyTorch. Install the CUDA-enabled PyTorch wheel for your platform from pytorch.org before installing PyPIELM.

# Example: CUDA 12.1
pip install torch --index-url https://download.pytorch.org/whl/cu121
pip install "pypielm[viz]"

On Apple Silicon, MPS acceleration is available out of the box with device="mps". Set PYTORCH_ENABLE_MPS_FALLBACK=1 to fall back to CPU for unsupported ops.

Verify Installation

import pypielm
print(pypielm.__version__)