pathak22/noreward-rl

Health v2 · maintenance only

Flags: no_releases no_license

Adoption not part of the score

1482 stars · 304 forks observed · 2026-08-28

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

TensorFlow implementation of the ICML 2017 paper 'Curiosity-driven Exploration by Self-supervised Prediction', training RL agents with an intrinsic curiosity module (ICM) for sparse-reward environments like Mario and VizDoom. It includes training and demo scripts plus pretrained models, supporting 'RL without rewards' exploration.

Use cases

• train RL agents in sparse-reward environments
• reproduce curiosity-driven exploration research
• run RL without external rewards using intrinsic motivation
• experiment with self-supervised prediction in deep RL
• benchmark exploration on Super Mario Bros and VizDoom

When to choose

• you need a reference implementation of the ICM curiosity paper
• your RL environment has sparse or no external rewards
• you want pretrained curiosity models for Mario or Doom

When to avoid

• you need a maintained framework for modern RL research
• you prefer PyTorch or current RL libraries
• you need production-grade, well-supported code

Facets

library · maturity maintenance

reinforcement-learning machine-learning deep-learning reinforcement-learning deep-learning machine-learning python curiosity-driven-exploration intrinsic-motivation self-supervised-learning tensorflow openai-gym icml-2017 research-code sparse-rewards linux gpu

Member repositories

For agents

markdown · JSON · MCP: product_card(name="pathak22/noreward-rl")