Beyond the Teacher : Novel pipeline for Imitation Learning (IL) Method

This repository contains the code, datasets, experiments, and evaluations for a proposed method in *Beyond the Teacher.

Folder	Description
Demo_utility/	Core Jupyter notebook for demo utility scoring and demo cleaning pipeline.
Picknplace/	Implements the Pick-and-Place task, including datasets, models, and scoring notebooks (evaluation scripts).
Weaving/	Implements the Weaving task, including datasets, models, and scoring notebooks (evaluation scripts)
Wiping/	Implements the Wiping task, including datasets, models, and scoring notebooks (evaluation scripts)

---

Folder Structure

1. Demo_utility/

Contains the core Jupyter notebook demo_utility.ipynb, which includes:

• Demo Utility Scoring – Quantitative evaluation of dataset usefulness for model training.
• Data Cleaning – Implements the proposed data-cleaning pipeline.

2. Picknplace/

Implements the Pick-and-Place task:

• data/ – Datasets for training and reward calculation.
• models/ – Jupyter notebooks for training IL models and generating rollouts.
• scoring/ – Notebooks for computing value maps and evaluating IL methods.

3. Weaving/

Implements the Weaving task:

• data/ – Task-specific datasets.
• models/ – Jupyter notebooks for model training and rollout generation.
• scoring/ – Notebooks for evaluation and scoring of IL methods.

4. Wiping/

Implements the Wiping task:

• data/ – Task-specific datasets.
• models/ – Jupyter notebooks for training and rollout generation.
• scoring/ – Notebooks for evaluation and scoring of IL methods.

---