MatthieuCourbariaux/BinaryNet

Health v2 · maintenance only

Flags: no_releases

Adoption not part of the score

1068 stars · 339 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

BinaryNet is a research codebase for training deep neural networks whose weights and activations are constrained to +1 or -1, reproducing the experiments from the BinaryNet paper. It includes train-time benchmark code and run-time XNOR/baseline GPU kernels.

Use cases

• train binary neural networks with +1/-1 weights and activations
• reproduce BinaryNet paper benchmark results
• run XNOR GPU kernels for fast binary network inference
• experiment with extremely low-bitwidth deep learning models
• study efficient neural network training techniques

When to choose

• you need to reproduce or extend the BinaryNet paper's experiments
• you are researching binary/quantized neural networks
• you want reference XNOR GPU kernel implementations

When to avoid

• you need a maintained production deep learning framework
• you want modern quantization support in PyTorch or TensorFlow
• you need up-to-date GPU compatibility and support

Facets

library · maturity maintenance

deep-learning machine-learning llm-training deep-learning machine-learning artificial-intelligence python binary-neural-networks bnn xnor-kernels research-code model-compression theano gpu

Member repositories

For agents

markdown · JSON · MCP: product_card(name="MatthieuCourbariaux/BinaryNet")