| # Model Card for Mozart vs Beethoven Classifier | |
| This model predicts whether a classical piano piece was composed by **Mozart** or **Beethoven**, based on numerical features extracted from the score (counts of right-hand notes, left-hand notes, measures, key centers, and markings). | |
| ## Model Details | |
| ### Model Description | |
| - **Developed by:** Scotty McGee (PhD student, CMU) | |
| - **Shared by [optional]:** Scotty McGee | |
| - **Model type:** Tabular classification (AutoML with AutoGluon) | |
| - **Language(s) (NLP):** Not applicable (tabular/numeric features only) | |
| - **License:** MIT (update as needed) | |
| - **Finetuned from model:** N/A | |
| ## Uses | |
| ### Direct Use | |
| Demonstration of machine-learning classification on musical data. Predicts a binary composer label (Mozart or Beethoven) from numeric score features. | |
| ### Downstream Use [optional] | |
| Could be adapted for broader composer classification tasks, musicology studies, or automated metadata tagging. | |
| ### Out-of-Scope Use | |
| - Not intended as a general music recognition tool. | |
| - Not reliable for real performance or music audio classification. | |
| - Not suitable for commercial music rights enforcement. | |
| ## Bias, Risks, and Limitations | |
| - Limited to Mozart and Beethoven; not generalizable to other composers. | |
| - Features are simplistic (counts of notes, measures, key centers, markings). | |
| - May not capture stylistic nuance. | |
| - Risk of overfitting to the dataset used. | |
| ### Recommendations | |
| Use for small-scale experiments and demos. Do not apply to large-scale music classification tasks without retraining and validation. | |
| ## How to Get Started with the Model | |
| ```python | |
| from autogluon.tabular import TabularPredictor | |
| preds = predictor.predict(df_test) # Mozart or Beethoven | |