Link model to paper and improve model card

Hi! I'm Niels from the community science team at Hugging Face.

This PR improves the model card for TabDPT by:
- Linking it to the original research paper via the `arxiv` metadata tag.
- Adding direct links to the official GitHub repository for inference and documentation.
- Updating the metadata with the `pipeline_tag: other` as requested.
- Organizing the README content to include installation and usage guidance based on the official repository.

README.md

@@ -1,24 +1,39 @@
 ---
-pipeline_tag: tabular-classification
+license: apache-2.0
+license_name: modified-mit
+pipeline_tag: other
 tags:
 - Tabular
 - In-Context-Learning
 - Transformer
-license_name: modified-mit
-license: apache-2.0
+arxiv: 2410.18164
 ---
-# TabDPT: Scaling Tabular Foundation Models
+
+# TabDPT: Scaling Tabular Foundation Models on Real Data
+
+**TabDPT** is an open-source foundation model for tabular data based on in-context learning (ICL). It is trained on real-world data and can generalize to new tasks across classification and regression without additional training or hyperparameter tuning.
+
+- **Paper:** [TabDPT: Scaling Tabular Foundation Models on Real Data](https://huggingface.co/papers/2410.18164)
+- **GitHub:** [layer6ai-labs/TabDPT](https://github.com/layer6ai-labs/TabDPT)
 
 ## Installation
-```
-git clone git@github.com:layer6ai-labs/TabDPT.git
+
+To set up the environment, ensure you have Python 3.10 or 3.11, then run:
+
+```bash
+git clone https://github.com/layer6ai-labs/TabDPT.git
 cd TabDPT
 pip install -e .
 ```
 
 ## Example Usage 
-Please refer to the GitHub repo.
-For better performance, please increase `context_size` or increase `n_ensembles` to trade off speed and accuracy
+
+TabDPT performs zero-shot prediction on new tabular datasets. For detailed working examples, please refer to the following files in the GitHub repository:
+- `tests/cls_example.py` for classification tasks.
+- `tests/reg_example.py` for regression tasks.
+
+**Performance Tips:**
+For better performance, you can increase `context_size` or increase `n_ensembles` to trade off speed and accuracy.
 
 ## Updates
 
@@ -29,11 +44,16 @@ For better performance, please increase `context_size` or increase `n_ensembles`
 Added support for flash attention (with bf16 precision) and compile flag. Both are enabled to True by default and should lead to a significant speed-up.
 
 ## Citation
-```
-@article{ma2024tabdpt,
-  title={TabDPT: Scaling Tabular Foundation Models},
-  author={Ma, Junwei and Thomas, Valentin and Hosseinzadeh, Rasa and Kamkari, Hamidreza and Labach, Alex and Cresswell, Jesse C and Golestan, Keyvan and Yu, Guangwei and Volkovs, Maksims and Caterini, Anthony L},
-  journal={arXiv preprint arXiv:2410.18164},
-  year={2024}
+
+```bibtex
+@inproceedings{
+    ma2025tabdpt,
+    title={Tab{DPT}: Scaling Tabular Foundation Models on Real Data},
+    author={Junwei Ma and Valentin Thomas and Rasa Hosseinzadeh and Alex Labach and Hamidreza Kamkari and Jesse C. Cresswell and Keyvan Golestan and Guangwei Yu and Anthony L. Caterini and Maksims Volkovs},
+    booktitle={The Thirty-ninth Annual Conference on Neural Information Processing Systems},
+    year={2025},
+    url={https://openreview.net/forum?id=pIZxEOZCId}
 }
-```
+```
+
+© Copyright 2024-2025 The Toronto-Dominion Bank and/or its affiliates