Improve model card: Add pipeline tag, links, license, and usage info

This PR improves the model card by:
- Adding the `pipeline_tag: robotics` metadata, enabling discovery via the tag.
- Correcting the license metadata from `apache-2.0` to `mit` based on the GitHub repository's badge.
- Linking to the paper: [Much Ado About Noising: Dispelling the Myths of Generative Robotic Control](https://huggingface.co/papers/2512.01809).
- Adding a link to the official project page: https://simchowitzlabpublic.github.io/much-ado-about-noising-project/.
- Adding a link to the GitHub repository: https://github.com/simchowitzlabpublic/much-ado-about-noising.
- Including key features and a "Quick Start" section with installation, training, and evaluation commands directly from the GitHub README for easier local setup and usage.
- Adding a basic BibTeX citation entry for the paper.

Please review and merge this PR if everything looks good.

README.md

@@ -1,3 +1,77 @@
----
-license: apache-2.0
----
+---
+license: mit
+pipeline_tag: robotics
+---
+
+# Much Ado About Noising: Dispelling the Myths of Generative Robotic Control
+
+This repository contains the code for the paper [Much Ado About Noising: Dispelling the Myths of Generative Robotic Control](https://huggingface.co/papers/2512.01809). It provides a PyTorch-based framework for behavior cloning with flow matching and related generative models, incorporating best practices from diffusion model training. The research investigates the factors underlying the success of generative control policies (GCPs) in robotics, suggesting their advantage stems from iterative computation with intermediate supervision and stochasticity, rather than multimodal action distributions or complex observation-to-action mappings.
+
+- **Project Page:** https://simchowitzlabpublic.github.io/much-ado-about-noising-project/
+- **Code Repository:** https://github.com/simchowitzlabpublic/much-ado-about-noising
+
+## Features
+
+- 🧩 **Clean & Modular**: Composable components for losses, samplers, networks, and encoders
+- ⚡ **Fast**: Optimized with torch.compile and CUDA graphs for maximum throughput
+- 📊 **Best Practices**: EMA, warmup scheduling, auto-resume, and proven training techniques
+- 🎯 **Diverse Algorithms**: Support for flow matching, consistency models, shortcut models, and regression
+- 🤖 **Robot-Ready**: Pre-configured for Robomimic, Kitchen, and PushT tasks
+
+## Installation
+
+```bash
+uv sync
+# install robomimic dependencies
+uv sync --extras robomimic
+# install kitchen dependencies
+uv sync --extras kitchen
+```
+
+## Quick Start
+
+### Training
+
+```bash
+# Train Robomimic (state observations)
+uv run examples/train_robomimic.py \
+    task=lift_ph_state \
+    network=chiunet \
+    optimization.loss_type=flow \
+    log.wandb_mode=online
+
+# Train Robomimic (image observations)
+uv run examples/train_robomimic.py \
+    task=lift_ph_image \
+    network=chiunet \
+    optimization.batch_size=256
+
+# Train Kitchen
+uv run examples/train_kitchen.py task=kitchen_state
+
+# Train PushT
+uv run examples/train_pusht.py task=pusht_state
+```
+
+### Evaluation
+
+```bash
+# Evaluate trained model
+uv run examples/train_robomimic.py \
+    mode=eval \
+    optimization.model_path="/path/to/checkpoint.pt"
+```
+
+## Citation
+
+If you find our work helpful, please cite the paper:
+
+```bibtex
+@article{anon2025much,
+  title={Much Ado About Noising: Dispelling the Myths of Generative Robotic Control},
+  author={Anonymous},
+  journal={arXiv preprint arXiv:2512.01809},
+  year={2025},
+  url={https://arxiv.org/abs/2512.01809}
+}
+```