carpedm20/ENAS-pytorch

Health v2 · maintenance only

Flags: no_releases

Adoption not part of the score

2719 stars · 486 forks observed · 2026-08-28

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

A PyTorch implementation of Efficient Neural Architecture Search (ENAS), which reduces NAS GPU-hours by 1000x via parameter sharing between subgraph models. It supports discovering recurrent cells for language modeling and CNN architectures, though the README recommends the authors' official code.

Use cases

• search for novel RNN cells for language modeling
• discover CNN architectures on CIFAR
• reproduce the ENAS paper in PyTorch
• run neural architecture search with limited GPU budget
• visualize discovered architectures as gifs

When to choose

• you want a PyTorch-based ENAS reference implementation
• you're studying or extending neural architecture search with parameter sharing

When to avoid

• you need production-grade or officially supported NAS code (use the authors' official implementation)
• you need modern PyTorch versions (it targets PyTorch 0.3.1)
• you need complete CNN search support (marked in progress)

Facets

library · maturity maintenance

machine-learning deep-learning machine-learning deep-learning python neural-architecture-search pytorch enas research-code parameter-sharing

Member repositories

For agents

markdown · JSON · MCP: product_card(name="carpedm20/ENAS-pytorch")