ag-behavior-cloning

Behavior cloning (imitation)

A policy trained by supervised imitation of expert demonstrations; 100% rollout success.

Trained from scratch in Ropedia Academy — an interactive, bilingual course on embodied & spatial AI. Educational model: small and quick to train; the value is the method and a reproducible pipeline, not a leaderboard score. Try it live in the Ropedia demos Space.

At a glance

Dataset

• Name: Expert gridworld demos
• Type: synthetic — procedural
• Size / stats: ~2,000 (state → expert action) pairs on a 6×6 grid (greedy expert)
• Split: train; eval = rollout success from every cell
• Source: procedural

Training config

Adam (lr 3e-3), 800 steps; cross-entropy on ~2k expert (state→action) pairs; 6×6 grid.

Evaluation results

Inference example

import torch
state = torch.load("policy.pt", map_location="cpu")   # this repo's checkpoint
# Rebuild the exact module from the lab notebook (see "Reproduce"), then:
# model.load_state_dict(state); model.eval()

Limitations

Educational scale. Trained quickly on CPU on small or synthetic data, so absolute numbers are not competitive with production systems — the value is the method and a reproducible pipeline. No large-scale data, no hyperparameter sweep, and no multi-seed variance is reported. Not for production use.

Failure cases

Compounding error / distribution shift once it leaves the expert's states (no recovery) — needs DAgger to fix.

Reproduce / train your own

One click: open the notebook in Colab → Runtime → GPU → Run all, then run its Publish to the Hugging Face Hub cell.

From a shell:

git clone https://github.com/ChaoYue0307/ropedia-academy.git && cd ropedia-academy
pip install torch numpy matplotlib scikit-learn scikit-image gymnasium
jupyter nbconvert --to notebook --execute notebooks/training/AG_behavior_cloning.ipynb --output run.ipynb
# optional: override training length, e.g.  STEPS=2000  (or EPISODES=600)  before running

Files

• figure.png
• metrics.json
• policy.pt

License

Code & weights: MIT (this repository) — educational use encouraged.
Data: generated procedurally in the notebook — no external dataset.

Citation

If you use this model or the course materials, please cite:

@misc{ropedia_academy,
  title  = {Ropedia Academy: an interactive course on embodied & spatial AI},
  author = {Ropedia Academy},
  year   = {2026},
  howpublished = {\url{https://chaoyue0307.github.io/ropedia-academy/}}
}

Method / original work: Pomerleau, ALVINN, 1988; Ross et al., DAgger, AISTATS 2011.

Related assets

• 🚀 Live demos: https://huggingface.co/spaces/cy0307/ropedia-demos
• 🤗 All trained models + collection: https://huggingface.co/cy0307
• 📚 Course & all labs: https://chaoyue0307.github.io/ropedia-academy/ · Labs tab
• 💻 Source / notebooks: github.com/ChaoYue0307/ropedia-academy

---

Part of the Ropedia Academy trained-model collection. Contributions & issues welcome on GitHub.