Upload README.md with huggingface_hub

README.md

@@ -1,32 +1,183 @@
 ---
 license: apache-2.0
+task_categories:
+  - question-answering
+  - visual-question-answering
+language:
+  - en
+tags:
+  - music
+  - music-understanding
+  - sheet-music
+  - abc-notation
+  - benchmark
+  - multimodal
+pretty_name: "MSU-Bench: Musical Score Understanding Benchmark"
+size_categories:
+  - n<1K
 configs:
-- config_name: default
-  data_files:
-  - split: test
-    path: data/test-*
+  - config_name: default
+    data_files:
+      - split: test
+        path: data/test-*
 dataset_info:
   features:
-  - name: song_id
-    dtype: string
-  - name: abc_notation
-    dtype: string
-  - name: pdf
-    dtype: binary
-  - name: images
-    list: image
-  - name: questions
-    struct:
-    - name: level
-      list: int32
-    - name: question
-      list: string
-    - name: answer
-      list: string
+    - name: song_id
+      dtype: string
+    - name: abc_notation
+      dtype: string
+    - name: pdf
+      dtype: binary
+    - name: images
+      sequence: image
+    - name: questions
+      struct:
+        - name: level
+          sequence: int32
+        - name: question
+          sequence: string
+        - name: answer
+          sequence: string
   splits:
-  - name: test
-    num_bytes: 209552791
-    num_examples: 150
+    - name: test
+      num_bytes: 209552791
+      num_examples: 150
   download_size: 207988689
   dataset_size: 209552791
 ---
+
+# MSU-Bench: Musical Score Understanding Benchmark
+
+[![arXiv](https://img.shields.io/badge/arXiv-2511.20697-b31b1b.svg)](https://arxiv.org/abs/2511.20697)
+[![HuggingFace](https://img.shields.io/badge/HuggingFace-Dataset-yellow.svg)](https://huggingface.co/datasets/Krinos/MSU-Bench)
+
+**Evaluating Large Language Models' Comprehension of Complete Musical Scores**
+
+> Accepted to **ACL 2026 Main Conference**
+
+---
+
+## Overview
+
+MSU-Bench is a human-curated benchmark for evaluating the musical score understanding capabilities of Large Language Models (LLMs) and Vision-Language Models (VLMs). It supports multimodal evaluation through both textual (ABC notation) and visual (PDF/image) inputs.
+
+**Key Statistics:**
+- 150 complete musical scores
+- 1,800 generative question-answer pairs
+- 4 hierarchical difficulty levels
+- 12 questions per score (3 per level)
+
+## Multi-level Understanding
+
+![ABC notation example](figures/abc.png)
+
+## Data Curation and Evaluation
+
+![Data curation pipeline](figures/data.png)
+
+## Difficulty Levels
+
+| Level | Focus | Examples |
+|-------|-------|----------|
+| **Level 1** - Onset Information | Metadata at the beginning of a score | Composer, title, key, time signature, tempo, instrumentation, anacrusis |
+| **Level 2** - Notation & Note | Local bar-level notation details | Pitch range, accidentals, dynamics, articulations, ornaments, tempo changes |
+| **Level 3** - Chord & Harmony | Harmonic structures and progressions | Chord qualities, inversions, cadences, modulations, pedal points |
+| **Level 4** - Texture & Form | Large-scale structural analysis | Melodic motifs, thematic organisation, texture types, formal design |
+
+## Dataset Structure
+
+Each of the 150 samples contains:
+
+| Column | Type | Description |
+|--------|------|-------------|
+| `song_id` | `string` | Unique identifier derived from the score filename |
+| `abc_notation` | `string` | Full ABC notation of the score (text-based symbolic representation) |
+| `pdf` | `binary` | The original rendered PDF of the score |
+| `images` | `list[image]` | Individual page images (PNG) of the score (1-35 pages) |
+| `questions` | `struct{level, question, answer}` | 12 questions per score (3 per difficulty level) with reference answers |
+
+### Modalities for Evaluation
+
+| Modality | Input | Target Models |
+|----------|-------|---------------|
+| Textual QA | ABC notation + question | LLMs |
+| Visual QA | PDF/images + question | VLMs |
+
+## Repertoire
+
+The benchmark covers 150 scores from the Western art music canon, spanning:
+- **Periods:** Baroque, Classical, Romantic, Impressionism, 20th Century
+- **Composers:** Bach, Beethoven, Chopin, Brahms, Debussy, Liszt, Schubert, Mozart, Mussorgsky, Grieg, and others
+- **Genres:** Sonatas, character pieces, fugues, waltzes, nocturnes, etudes, rhapsodies, symphonies, concertos, art songs
+
+## Usage
+
+### Loading the Dataset
+
+```python
+from datasets import load_dataset
+
+ds = load_dataset("Krinos/MSU-Bench", split="test")
+
+# Access a sample
+sample = ds[0]
+print(sample["song_id"])        # '1._Gnomus_The_Gnome__Promenade_...'
+print(sample["abc_notation"][:200])
+print(len(sample["images"]))    # number of page images
+print(sample["questions"])      # {'level': [...], 'question': [...], 'answer': [...]}
+```
+
+### Iterating Over Questions
+
+```python
+# Flatten to one row per question for evaluation
+flat_data = []
+for sample in ds:
+    levels = sample["questions"]["level"]
+    questions = sample["questions"]["question"]
+    answers = sample["questions"]["answer"]
+    for lvl, q, a in zip(levels, questions, answers):
+        flat_data.append({
+            "song_id": sample["song_id"],
+            "level": lvl,
+            "question": q,
+            "answer": a,
+            "abc_notation": sample["abc_notation"],
+        })
+
+print(f"Total QA pairs: {len(flat_data)}")  # 1800
+```
+
+### Visual QA with Page Images
+
+```python
+# Get page images for a score
+sample = ds[0]
+for i, img in enumerate(sample["images"]):
+    img.save(f"page_{i}.png")
+```
+
+## Evaluation Protocol
+
+We adopt an **LLM-as-a-judge** framework using majority voting across three models (ChatGPT-5, Claude Sonnet 4, Gemini 2.5 Pro) to evaluate semantic correctness. This approach handles equivalent musical terminology (e.g., "V-I" vs. "authentic cadence") better than strict string matching.
+
+**Inter-judge agreement:** 83.33% (95% CI: 79.44%-87.22%)
+
+## Citation
+
+```bibtex
+@article{dai2025msubench,
+  title={Musical Score Understanding Benchmark: Evaluating Large Language Models' Comprehension of Complete Musical Scores},
+  author={Dai, Congren and Yang, Yue and Li, Krinos and Zhou, Huichi and Liang, Shijie and Zhang, Bo and Liu, Enyang and Jin, Ge and An, Hongran and Zhang, Haosen and Jing, Peiyuan and Lee, Kinhei and Zhang, Zhenxuan and Li, Xiaobing and Sun, Maosong},
+  journal={arXiv preprint arXiv:2511.20697},
+  year={2025}
+}
+```
+
+## License
+
+Apache 2.0. Please also refer to the repository for licensing information regarding the musical scores sourced from MuseScore.
+
+## Acknowledgements
+
+This work is supported by the Advanced Discipline Construction Project of Beijing Universities, the Special Programme of National Natural Science Foundation of China (Grant No. T2341003), and the Major Programme of National Social Science Fund of China (Grant No. 21ZD19).