Hyperparameter Optimization and Reproducibility in Deep Learning Model Training
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866915589176426496 |
|---|---|
| author | Afzaal, Usman Su, Ziyu Sajjad, Usama Lu, Hao Rezapour, Mostafa Gurcan, Metin Nafi Niazi, Muhammad Khalid Khan |
| author_facet | Afzaal, Usman Su, Ziyu Sajjad, Usama Lu, Hao Rezapour, Mostafa Gurcan, Metin Nafi Niazi, Muhammad Khalid Khan |
| contents | Reproducibility remains a critical challenge in foundation model training for histopathology, often hindered by software randomness, hardware non-determinism, and inconsistent hyperparameter reporting. To investigate these issues, we trained a CLIP model on the QUILT-1M dataset and systematically evaluated the impact of different hyperparameter settings and augmentation strategies across three downstream histopathology datasets (PatchCamelyon, LC25000-Lung, and LC25000-Colon). Despite variability across runs, we identified clear trends: RandomResizedCrop values of 0.7-0.8 outperformed more aggressive (0.6) or conservative (0.9) settings, distributed training without local loss improved stability, and learning rates below 5.0e-5 consistently degraded performance across all datasets. The LC25000 (Colon) dataset consistently provided the most reproducible benchmark. These findings highlight that reproducibility in computational pathology depends not only on transparent documentation but also on carefully chosen experimental configurations, and we provide practical rules to guide future efforts in developing reproducible foundation models for digital pathology. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2510_15164 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Hyperparameter Optimization and Reproducibility in Deep Learning Model Training Afzaal, Usman Su, Ziyu Sajjad, Usama Lu, Hao Rezapour, Mostafa Gurcan, Metin Nafi Niazi, Muhammad Khalid Khan Computer Vision and Pattern Recognition Reproducibility remains a critical challenge in foundation model training for histopathology, often hindered by software randomness, hardware non-determinism, and inconsistent hyperparameter reporting. To investigate these issues, we trained a CLIP model on the QUILT-1M dataset and systematically evaluated the impact of different hyperparameter settings and augmentation strategies across three downstream histopathology datasets (PatchCamelyon, LC25000-Lung, and LC25000-Colon). Despite variability across runs, we identified clear trends: RandomResizedCrop values of 0.7-0.8 outperformed more aggressive (0.6) or conservative (0.9) settings, distributed training without local loss improved stability, and learning rates below 5.0e-5 consistently degraded performance across all datasets. The LC25000 (Colon) dataset consistently provided the most reproducible benchmark. These findings highlight that reproducibility in computational pathology depends not only on transparent documentation but also on carefully chosen experimental configurations, and we provide practical rules to guide future efforts in developing reproducible foundation models for digital pathology. |
| title | Hyperparameter Optimization and Reproducibility in Deep Learning Model Training |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2510.15164 |