Installation Guide

Installation Guide#

Requirements#

  • Python 3.8 or higher

  • PyTorch 1.10 or higher

  • NumPy, Pandas, SciPy

Install from PyPI#

pip install corerec

Install from Source#

git clone https://github.com/vishesh9131/CoreRec.git
cd CoreRec
pip install -e .

Optional Extras#

# Serving (FastAPI REST API)
pip install "corerec[serving]"

# Tutorial datasets (cr_learn)
pip install "corerec[datasets]"

# Development + tests
pip install "corerec[dev]"

# Everything
pip install "corerec[all]"

See also: CoreRec API Versioning Policy (API stability policy) and Vendored torch_nn and torch_utils (internal PyTorch modules).

Install cr_learn (for tutorials)#

Tutorial examples use the cr_learn dataset package:

pip install cr_learn

The first run of ml_1m.load() downloads MovieLens 1M (~25 MB) to your local cache.

Environment Notes#

  • NumPy / PyTorch: If you see NumPy compatibility warnings with PyTorch, use a matched pair, e.g. pip install 'numpy<2' with older PyTorch builds, or upgrade PyTorch to a NumPy 2–compatible release.

  • GPU: Deep learning models auto-detect CUDA when available; CPU works for small examples.

Verify Installation#

import corerec
print(corerec.__version__)

# Test import
from corerec.engines.dcn import DCN
model = DCN()
print("✅ Installation successful!")

Troubleshooting#

CUDA Issues#

If you encounter CUDA errors:

pip install torch torchvision torchaudio --extra-index-url https://download.pytorch.org/whl/cu116

ImportError#

If you get import errors after installation:

pip install --upgrade corerec
Contents