Add paper link, GitHub repository, and model metadata
Browse filesHi! I'm Niels from the community science team at Hugging Face.
I've opened this PR to improve the model card for TAMA-QWen3. This update includes:
- Structured metadata for better discoverability (`pipeline_tag` and `library_name`).
- Prominent links to the [MMTU paper page](https://huggingface.co/papers/2506.05587) and the [official GitHub repository](https://github.com/mmtu-benchmark/mmtu).
- A BibTeX citation section for the benchmark.
- Preservation of your original performance table and styling.
These changes help users find the model through Hub filtering and provide clear attribution to the research.
README.md
CHANGED
|
@@ -1,5 +1,17 @@
|
|
| 1 |
---
|
| 2 |
license: apache-2.0
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 3 |
---
|
| 4 |
|
| 5 |
<!doctype html>
|
|
@@ -198,7 +210,7 @@ license: apache-2.0
|
|
| 198 |
<br/></td><td id='temp:s:temp:C:GJGa47d78fca9d44923919996e67;temp:C:GJGbb5cc10122964720bc8da76c5' style='background-color:#72c69d;text-align: right;vertical-align: bottom;' class="extra-columns">73.7
|
| 199 |
<br/></td><td id='temp:s:temp:C:GJGa47d78fca9d44923919996e67;temp:C:GJG0d0b379bcb6f462185bacedaa' style='background-color:#b2e0ca;text-align: right;vertical-align: bottom;' class="extra-columns">14.3
|
| 200 |
<br/></td><td id='temp:s:temp:C:GJGa47d78fca9d44923919996e67;temp:C:GJG750f96b656904402b0685a9dc' style='background-color:#a5dbc1;text-align: right;vertical-align: bottom;' class="extra-columns">20.9
|
| 201 |
-
<br/></td><td id='temp:s:temp:C:GJGa47d78fca9d44923919996e67;temp:C:
|
| 202 |
<br/></td><td id='temp:s:temp:C:GJGa47d78fca9d44923919996e67;temp:C:GJGdbb0b3f228a24416873a958a9' style='background-color:#bde4d1;text-align: right;vertical-align: bottom;' class="extra-columns">5.9
|
| 203 |
<br/></td><td id='temp:s:temp:C:GJGa47d78fca9d44923919996e67;temp:C:GJG6886327be27f435c846f58863' style='background-color:#e3f4ec;text-align: right;vertical-align: bottom;' class="extra-columns">0.1
|
| 204 |
<br/></td></tr><tr id='temp:C:GJG49cecd182f02483f9e1a67399'><td id='temp:s:temp:C:GJG49cecd182f02483f9e1a67399;temp:C:GJG35c2ff53e6ca4ff6a432d6e10' style=''><a href="https://huggingface.co/MichiganNLP/TAMA-vA">TAMA-vA</a>
|
|
@@ -212,7 +224,7 @@ license: apache-2.0
|
|
| 212 |
<br/></td><td id='temp:s:temp:C:GJG49cecd182f02483f9e1a67399;temp:C:GJGbb5cc10122964720bc8da76c5' style='background-color:#afdfc7;text-align: right;vertical-align: bottom;' class="extra-columns">41.8
|
| 213 |
<br/></td><td id='temp:s:temp:C:GJG49cecd182f02483f9e1a67399;temp:C:GJG0d0b379bcb6f462185bacedaa' style='background-color:#bbe4d0;text-align: right;vertical-align: bottom;' class="extra-columns">12.7
|
| 214 |
<br/></td><td id='temp:s:temp:C:GJG49cecd182f02483f9e1a67399;temp:C:GJG750f96b656904402b0685a9dc' style='background-color:#a4dbc0;text-align: right;vertical-align: bottom;' class="extra-columns">21.1
|
| 215 |
-
<br/></td><td id='temp:s:temp:C:GJG49cecd182f02483f9e1a67399;temp:C:
|
| 216 |
<br/></td><td id='temp:s:temp:C:GJG49cecd182f02483f9e1a67399;temp:C:GJGdbb0b3f228a24416873a958a9' style='background-color:#b8e3ce;text-align: right;vertical-align: bottom;' class="extra-columns">6.3
|
| 217 |
<br/></td><td id='temp:s:temp:C:GJG49cecd182f02483f9e1a67399;temp:C:GJG6886327be27f435c846f58863' style='background-color:#e3f4ec;text-align: right;vertical-align: bottom;' class="extra-columns">0.1
|
| 218 |
<br/></td></tr><tr id='temp:C:GJGb7a26b478b8a4d66bce379049'><td id='temp:s:temp:C:GJGb7a26b478b8a4d66bce379049;temp:C:GJG35c2ff53e6ca4ff6a432d6e10' style=''><a href="https://huggingface.co/Qwen/Qwen2.5-7B-Instruct">Qwen2.5-7B</a>
|
|
@@ -282,7 +294,7 @@ license: apache-2.0
|
|
| 282 |
<br/></td><td id='temp:s:temp:C:GJG0085c5b822744c81b434cc6cd;temp:C:GJGbb5cc10122964720bc8da76c5' style='background-color:#5fbf90;text-align: right;vertical-align: bottom;' class="extra-columns">83.2
|
| 283 |
<br/></td><td id='temp:s:temp:C:GJG0085c5b822744c81b434cc6cd;temp:C:GJG0d0b379bcb6f462185bacedaa' style='background-color:#63c093;text-align: right;vertical-align: bottom;' class="extra-columns">28.9
|
| 284 |
<br/></td><td id='temp:s:temp:C:GJG0085c5b822744c81b434cc6cd;temp:C:GJG750f96b656904402b0685a9dc' style='background-color:#57bb8a;text-align: right;vertical-align: bottom;' class="extra-columns">38.9
|
| 285 |
-
<br/></td><td id='temp:s:temp:C:GJG0085c5b822744c81b434cc6cd;temp:C:
|
| 286 |
<br/></td><td id='temp:s:temp:C:GJG0085c5b822744c81b434cc6cd;temp:C:GJGdbb0b3f228a24416873a958a9' style='background-color:#5ebe8f;text-align: right;vertical-align: bottom;' class="extra-columns">14.2
|
| 287 |
<br/></td><td id='temp:s:temp:C:GJG0085c5b822744c81b434cc6cd;temp:C:GJG6886327be27f435c846f58863' style='background-color:#57bb8a;text-align: right;vertical-align: bottom;' class="extra-columns">0.6
|
| 288 |
<br/></td></tr><tr id='temp:C:GJG3cb5f57e8e3748a9b13fe5018'><td id='temp:s:temp:C:GJG3cb5f57e8e3748a9b13fe5018;temp:C:GJG35c2ff53e6ca4ff6a432d6e10' style=''><a href="https://huggingface.co/MichiganNLP/TAMA-QWen3">TAMA-QWen3</a>
|
|
@@ -321,3 +333,13 @@ license: apache-2.0
|
|
| 321 |
</body>
|
| 322 |
</html>
|
| 323 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
---
|
| 2 |
license: apache-2.0
|
| 3 |
+
library_name: transformers
|
| 4 |
+
pipeline_tag: table-question-answering
|
| 5 |
+
base_model: Qwen/Qwen3-8B
|
| 6 |
+
---
|
| 7 |
+
|
| 8 |
+
# TAMA-QWen3
|
| 9 |
+
|
| 10 |
+
This repository contains **TAMA-QWen3**, an instruction-tuned model based on [Qwen3-8B](https://huggingface.co/Qwen/Qwen3-8B) optimized for table understanding and reasoning. It was introduced and evaluated as part of the work: **[MMTU: A Massive Multi-Task Table Understanding and Reasoning Benchmark](https://huggingface.co/papers/2506.05587)**.
|
| 11 |
+
|
| 12 |
+
- **GitHub Repository:** [MMTU-Benchmark/MMTU](https://github.com/mmtu-benchmark/mmtu)
|
| 13 |
+
- **Paper:** [arXiv:2506.05587](https://arxiv.org/abs/2506.05587)
|
| 14 |
+
|
| 15 |
---
|
| 16 |
|
| 17 |
<!doctype html>
|
|
|
|
| 210 |
<br/></td><td id='temp:s:temp:C:GJGa47d78fca9d44923919996e67;temp:C:GJGbb5cc10122964720bc8da76c5' style='background-color:#72c69d;text-align: right;vertical-align: bottom;' class="extra-columns">73.7
|
| 211 |
<br/></td><td id='temp:s:temp:C:GJGa47d78fca9d44923919996e67;temp:C:GJG0d0b379bcb6f462185bacedaa' style='background-color:#b2e0ca;text-align: right;vertical-align: bottom;' class="extra-columns">14.3
|
| 212 |
<br/></td><td id='temp:s:temp:C:GJGa47d78fca9d44923919996e67;temp:C:GJG750f96b656904402b0685a9dc' style='background-color:#a5dbc1;text-align: right;vertical-align: bottom;' class="extra-columns">20.9
|
| 213 |
+
<br/></td><td id='temp:s:temp:C:GJGa47d78fca9d44923919996e67;temp:C:GJG090381cf461f4be6be1be545a3b6' style='background-color:#8ed1b0;text-align: right;vertical-align: bottom;' class="extra-columns">18.7
|
| 214 |
<br/></td><td id='temp:s:temp:C:GJGa47d78fca9d44923919996e67;temp:C:GJGdbb0b3f228a24416873a958a9' style='background-color:#bde4d1;text-align: right;vertical-align: bottom;' class="extra-columns">5.9
|
| 215 |
<br/></td><td id='temp:s:temp:C:GJGa47d78fca9d44923919996e67;temp:C:GJG6886327be27f435c846f58863' style='background-color:#e3f4ec;text-align: right;vertical-align: bottom;' class="extra-columns">0.1
|
| 216 |
<br/></td></tr><tr id='temp:C:GJG49cecd182f02483f9e1a67399'><td id='temp:s:temp:C:GJG49cecd182f02483f9e1a67399;temp:C:GJG35c2ff53e6ca4ff6a432d6e10' style=''><a href="https://huggingface.co/MichiganNLP/TAMA-vA">TAMA-vA</a>
|
|
|
|
| 224 |
<br/></td><td id='temp:s:temp:C:GJG49cecd182f02483f9e1a67399;temp:C:GJGbb5cc10122964720bc8da76c5' style='background-color:#afdfc7;text-align: right;vertical-align: bottom;' class="extra-columns">41.8
|
| 225 |
<br/></td><td id='temp:s:temp:C:GJG49cecd182f02483f9e1a67399;temp:C:GJG0d0b379bcb6f462185bacedaa' style='background-color:#bbe4d0;text-align: right;vertical-align: bottom;' class="extra-columns">12.7
|
| 226 |
<br/></td><td id='temp:s:temp:C:GJG49cecd182f02483f9e1a67399;temp:C:GJG750f96b656904402b0685a9dc' style='background-color:#a4dbc0;text-align: right;vertical-align: bottom;' class="extra-columns">21.1
|
| 227 |
+
<br/></td><td id='temp:s:temp:C:GJG49cecd182f02483f9e1a67399;temp:C:GJG090381cf461f4be6be1be545a3b6' style='background-color:#97d5b7;text-align: right;vertical-align: bottom;' class="extra-columns">17.1
|
| 228 |
<br/></td><td id='temp:s:temp:C:GJG49cecd182f02483f9e1a67399;temp:C:GJGdbb0b3f228a24416873a958a9' style='background-color:#b8e3ce;text-align: right;vertical-align: bottom;' class="extra-columns">6.3
|
| 229 |
<br/></td><td id='temp:s:temp:C:GJG49cecd182f02483f9e1a67399;temp:C:GJG6886327be27f435c846f58863' style='background-color:#e3f4ec;text-align: right;vertical-align: bottom;' class="extra-columns">0.1
|
| 230 |
<br/></td></tr><tr id='temp:C:GJGb7a26b478b8a4d66bce379049'><td id='temp:s:temp:C:GJGb7a26b478b8a4d66bce379049;temp:C:GJG35c2ff53e6ca4ff6a432d6e10' style=''><a href="https://huggingface.co/Qwen/Qwen2.5-7B-Instruct">Qwen2.5-7B</a>
|
|
|
|
| 294 |
<br/></td><td id='temp:s:temp:C:GJG0085c5b822744c81b434cc6cd;temp:C:GJGbb5cc10122964720bc8da76c5' style='background-color:#5fbf90;text-align: right;vertical-align: bottom;' class="extra-columns">83.2
|
| 295 |
<br/></td><td id='temp:s:temp:C:GJG0085c5b822744c81b434cc6cd;temp:C:GJG0d0b379bcb6f462185bacedaa' style='background-color:#63c093;text-align: right;vertical-align: bottom;' class="extra-columns">28.9
|
| 296 |
<br/></td><td id='temp:s:temp:C:GJG0085c5b822744c81b434cc6cd;temp:C:GJG750f96b656904402b0685a9dc' style='background-color:#57bb8a;text-align: right;vertical-align: bottom;' class="extra-columns">38.9
|
| 297 |
+
<br/></td><td id='temp:s:temp:C:GJG0085c5b822744c81b434cc6cd;temp:C:GJG090381cf461f4be6be1be545a3b6' style='background-color:#57bb8a;text-align: right;vertical-align: bottom;' class="extra-columns">27.6
|
| 298 |
<br/></td><td id='temp:s:temp:C:GJG0085c5b822744c81b434cc6cd;temp:C:GJGdbb0b3f228a24416873a958a9' style='background-color:#5ebe8f;text-align: right;vertical-align: bottom;' class="extra-columns">14.2
|
| 299 |
<br/></td><td id='temp:s:temp:C:GJG0085c5b822744c81b434cc6cd;temp:C:GJG6886327be27f435c846f58863' style='background-color:#57bb8a;text-align: right;vertical-align: bottom;' class="extra-columns">0.6
|
| 300 |
<br/></td></tr><tr id='temp:C:GJG3cb5f57e8e3748a9b13fe5018'><td id='temp:s:temp:C:GJG3cb5f57e8e3748a9b13fe5018;temp:C:GJG35c2ff53e6ca4ff6a432d6e10' style=''><a href="https://huggingface.co/MichiganNLP/TAMA-QWen3">TAMA-QWen3</a>
|
|
|
|
| 333 |
</body>
|
| 334 |
</html>
|
| 335 |
|
| 336 |
+
## Citation
|
| 337 |
+
|
| 338 |
+
```bibtex
|
| 339 |
+
@article{mmtu,
|
| 340 |
+
title={{MMTU}: A Massive Multi-Task Table Understanding and Reasoning Benchmark},
|
| 341 |
+
author={Junjie Xing and Yeye He and Mengyu Zhou and Haoyu Dong and Shi Han and Lingjiao Chen and Dongmei Zhang and Surajit Chaudhuri and H. V. Jagadish},
|
| 342 |
+
journal={arXiv preprint arXiv:2506.05587},
|
| 343 |
+
year={2025}
|
| 344 |
+
}
|
| 345 |
+
```
|