CTG GPU Benchmark Pipeline: Transformer-Based Fetal Monitoring Models with Reproducible Python Scripts
Fuente:
Zenodo
Guardado en:
| Autor principal: | |
|---|---|
| Formato: | Recurso digital |
| Lenguaje: | inglés |
| Publicado: |
Zenodo
2025
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866902250849304576 |
|---|---|
| author | Ferhat Karataş |
| author_facet | Ferhat Karataş |
| contents | <h3><strong>1. Overview</strong></h3> <p>This repository provides a reproducible GPU benchmark pipeline for transformer-based fetal monitoring models using cardiotocography (CTG) data.<br>It includes complete scripts for<strong> </strong>signal preprocessing, data segmentation, model training, and prediction scoring, enabling fair performance comparison across multiple deep learning architectures.</p> <p>All data are anonymized.<br>The codebase is optimized for GPU environments such as Google Colab, Kaggle Notebooks, or local CUDA setups, and can be executed interactively (via Jupyter) or automatically (via Python CLI).</p> <h3><strong>2. Repository Contents</strong></h3> <div> <div> <table> <thead> <tr> <th>File</th> <th>Type</th> <th>Description</th> </tr> </thead> <tbody> <tr> <td><strong>01_ExamData.csv</strong></td> <td>Raw Data file</td> <td>Sample dataset containing fetal monitoring recordings with three synchronized channels: <strong>FHR (Fetal Heart Rate)</strong>, <strong>UA (Uterine Activity)</strong>, and <strong>AFM (Automatic Fetal Movement)</strong>.</td> </tr> <tr> <td><strong>02_SplitData.ipynb</strong></td> <td>Jupyter Notebook</td> <td>Preprocessing and segmentation notebook that splits the raw data into training, validation, and test sets.</td> </tr> <tr> <td><strong>19_GPU.ipynb</strong></td> <td>Jupyter Notebook</td> <td>Interactive GPU training notebook for multiple time-series architectures (e.g., <strong>TSMixer, PatchTST, iTransformer, TCN, ResNet1D</strong>). Computes performance metrics such as MAE, RMSE, and R².</td> </tr> <tr> <td><strong>ctg_gpu_benchmark_v3_full.py</strong></td> <td>Python script</td> <td>Fully automated pipeline script for command-line execution. Handles model selection, dataset loading, GPU training loops, and output storage.</td> </tr> <tr> <td><strong>_scored_output.xlsx</strong></td> <td>Data file</td> <td>Contains model predictions and ground-truth values on the test set for subsequent statistical analysis and figure generation.</td> </tr> </tbody> </table> </div> </div> <h3><strong>3. Usage Notes</strong></h3> <p>Dependencies:</p> <ul> <li> <p>Python ≥ 3.10</p> </li> <li> <p>PyTorch ≥ 2.1</p> </li> <li> <p>scikit-learn, pandas, numpy, matplotlib, tqdm</p> </li> <li> <p>GPU with CUDA support recommended</p> </li> </ul> <p>Suggested Execution Order:<br>1️⃣ <code>01_ExamData.csv</code> → raw data from NST signals<br>2️⃣ <code>02_SplitData.ipynb</code> → preprocessing & segmentation<br>3️⃣ <code>19_GPU.ipynb</code> → model training (interactive mode)<br>4️⃣ <code>ctg_gpu_benchmark_v3_full.py</code> → automated benchmark<br>5️⃣ <code>_scored_output.xlsx</code> → source dataset</p> <p>License & Citation:<br>This dataset and code package are released for academic and research use.<br>When referencing, please cite as:</p> <blockquote> <p>Karataş S. et al., <em>“CTG GPU Benchmark Pipeline: Transformer-Based Fetal Monitoring Models with Reproducible Python Scripts,”</em> Zenodo (2025). DOI: <span>https://doi.org/10.5281/zenodo.17456942</span></p> </blockquote> <h3><strong>4. Purpose</strong></h3> <p>The goal of this project is to establish a transparent and reproducible benchmark for <strong>deep learning models applied to fetal cardiotocography (CTG)</strong>.<br>By standardizing preprocessing, segmentation, and evaluation procedures, this pipeline enables objective performance comparison across transformer and classical time-series models, supporting future research in <strong>clinical fetal health prediction and perinatal data science</strong>.</p> <h3><strong>5. Keywords</strong></h3> <p><code>Cardiotocography</code>, <code>Fetal Monitoring</code>, <code>Transformer Models</code>, <code>Deep Learning</code>, <code>GPU Benchmark</code>, <code>Reproducibility</code>, <code>Python</code>, <code>Time-Series Analysis</code></p> |
| format | Recurso digital |
| id | zenodo_https___doi_org_10_5281_zenodo_17456942 |
| institution | Zenodo |
| language | eng |
| publishDate | 2025 |
| publisher | Zenodo |
| record_format | zenodo |
| spellingShingle | CTG GPU Benchmark Pipeline: Transformer-Based Fetal Monitoring Models with Reproducible Python Scripts Ferhat Karataş Cardiotocography CTG fetal heart rate uterine activity AFM time series classification regression Transformers PatchTST iTransformer TimesNet <h3><strong>1. Overview</strong></h3> <p>This repository provides a reproducible GPU benchmark pipeline for transformer-based fetal monitoring models using cardiotocography (CTG) data.<br>It includes complete scripts for<strong> </strong>signal preprocessing, data segmentation, model training, and prediction scoring, enabling fair performance comparison across multiple deep learning architectures.</p> <p>All data are anonymized.<br>The codebase is optimized for GPU environments such as Google Colab, Kaggle Notebooks, or local CUDA setups, and can be executed interactively (via Jupyter) or automatically (via Python CLI).</p> <h3><strong>2. Repository Contents</strong></h3> <div> <div> <table> <thead> <tr> <th>File</th> <th>Type</th> <th>Description</th> </tr> </thead> <tbody> <tr> <td><strong>01_ExamData.csv</strong></td> <td>Raw Data file</td> <td>Sample dataset containing fetal monitoring recordings with three synchronized channels: <strong>FHR (Fetal Heart Rate)</strong>, <strong>UA (Uterine Activity)</strong>, and <strong>AFM (Automatic Fetal Movement)</strong>.</td> </tr> <tr> <td><strong>02_SplitData.ipynb</strong></td> <td>Jupyter Notebook</td> <td>Preprocessing and segmentation notebook that splits the raw data into training, validation, and test sets.</td> </tr> <tr> <td><strong>19_GPU.ipynb</strong></td> <td>Jupyter Notebook</td> <td>Interactive GPU training notebook for multiple time-series architectures (e.g., <strong>TSMixer, PatchTST, iTransformer, TCN, ResNet1D</strong>). Computes performance metrics such as MAE, RMSE, and R².</td> </tr> <tr> <td><strong>ctg_gpu_benchmark_v3_full.py</strong></td> <td>Python script</td> <td>Fully automated pipeline script for command-line execution. Handles model selection, dataset loading, GPU training loops, and output storage.</td> </tr> <tr> <td><strong>_scored_output.xlsx</strong></td> <td>Data file</td> <td>Contains model predictions and ground-truth values on the test set for subsequent statistical analysis and figure generation.</td> </tr> </tbody> </table> </div> </div> <h3><strong>3. Usage Notes</strong></h3> <p>Dependencies:</p> <ul> <li> <p>Python ≥ 3.10</p> </li> <li> <p>PyTorch ≥ 2.1</p> </li> <li> <p>scikit-learn, pandas, numpy, matplotlib, tqdm</p> </li> <li> <p>GPU with CUDA support recommended</p> </li> </ul> <p>Suggested Execution Order:<br>1️⃣ <code>01_ExamData.csv</code> → raw data from NST signals<br>2️⃣ <code>02_SplitData.ipynb</code> → preprocessing & segmentation<br>3️⃣ <code>19_GPU.ipynb</code> → model training (interactive mode)<br>4️⃣ <code>ctg_gpu_benchmark_v3_full.py</code> → automated benchmark<br>5️⃣ <code>_scored_output.xlsx</code> → source dataset</p> <p>License & Citation:<br>This dataset and code package are released for academic and research use.<br>When referencing, please cite as:</p> <blockquote> <p>Karataş S. et al., <em>“CTG GPU Benchmark Pipeline: Transformer-Based Fetal Monitoring Models with Reproducible Python Scripts,”</em> Zenodo (2025). DOI: <span>https://doi.org/10.5281/zenodo.17456942</span></p> </blockquote> <h3><strong>4. Purpose</strong></h3> <p>The goal of this project is to establish a transparent and reproducible benchmark for <strong>deep learning models applied to fetal cardiotocography (CTG)</strong>.<br>By standardizing preprocessing, segmentation, and evaluation procedures, this pipeline enables objective performance comparison across transformer and classical time-series models, supporting future research in <strong>clinical fetal health prediction and perinatal data science</strong>.</p> <h3><strong>5. Keywords</strong></h3> <p><code>Cardiotocography</code>, <code>Fetal Monitoring</code>, <code>Transformer Models</code>, <code>Deep Learning</code>, <code>GPU Benchmark</code>, <code>Reproducibility</code>, <code>Python</code>, <code>Time-Series Analysis</code></p> |
| title | CTG GPU Benchmark Pipeline: Transformer-Based Fetal Monitoring Models with Reproducible Python Scripts |
| topic | Cardiotocography CTG fetal heart rate uterine activity AFM time series classification regression Transformers PatchTST iTransformer TimesNet |
| url | https://doi.org/10.5281/zenodo.17456942 |