BERTology of Molecular Property Prediction (Cross-validated Hyperparameter Search: Base-BERT)

Fuente: Zenodo
Guardado en:
Detalles Bibliográficos
Autores principales: Mostafanejad, Mohammad, Saxe, Paul, Crawford, T. Daniel
Formato: Recurso digital
Lenguaje:inglés
Publicado: Zenodo 2026
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866901640704950272
author Mostafanejad, Mohammad
Saxe, Paul
Crawford, T. Daniel
author_facet Mostafanejad, Mohammad
Saxe, Paul
Crawford, T. Daniel
contents <div> <h1>Cross-validated Hyperparameter Search</h1> <h2>Base-BERT</h2> This record pertains to Cross-validated Hyperparameter Search Experiments involving Base-BERT.</div> <div>It includes the following files:</div> <div> <ul> <li>HLM-fold0-base-binidx-0-ms-1234-ds-2345.tar.gz</li> <li>HLM-fold0-base-binidx-3-ms-1234-ds-1234.tar.gz</li> <li>HLM-fold0-base-binidx-5-ms-1234-ds-1234.tar.gz</li> <li>HLM-fold1-base-binidx-0-ms-1234-ds-2345.tar.gz</li> <li>HLM-fold1-base-binidx-3-ms-1234-ds-1234.tar.gz</li> <li>HLM-fold1-base-binidx-5-ms-1234-ds-1234.tar.gz</li> <li>HLM-fold2-base-binidx-0-ms-1234-ds-2345.tar.gz</li> <li>HLM-fold2-base-binidx-3-ms-1234-ds-1234.tar.gz</li> <li>HLM-fold2-base-binidx-5-ms-1234-ds-1234.tar.gz</li> <li>hPPB-fold0-base-binidx-0-ms-1234-ds-2345.tar.gz</li> <li>hPPB-fold0-base-binidx-3-ms-1234-ds-1234.tar.gz</li> <li>hPPB-fold0-base-binidx-5-ms-1234-ds-1234.tar.gz</li> <li>hPPB-fold1-base-binidx-0-ms-1234-ds-2345.tar.gz<br>hPPB-fold1-base-binidx-3-ms-1234-ds-1234.tar.gz</li> <li>hPPB-fold1-base-binidx-5-ms-1234-ds-1234.tar.gz</li> <li>hPPB-fold2-base-binidx-0-ms-1234-ds-2345.tar.gz</li> <li>hPPB-fold2-base-binidx-3-ms-1234-ds-1234.tar.gz</li> <li>hPPB-fold2-base-binidx-5-ms-1234-ds-1234.tar.gz</li> <li>MDR1_ER-fold0-base-binidx-5-ms-1234-ds-1234.tar.gz</li> <li>MDR1_ER-fold1-base-binidx-5-ms-1234-ds-1234.tar.gz</li> <li>MDR1_ER-fold2-base-binidx-5-ms-1234-ds-1234.tar.gz</li> <li>RLM-fold0-base-binidx-5-ms-1234-ds-1234.tar.gz</li> <li>RLM-fold1-base-binidx-5-ms-1234-ds-1234.tar.gz</li> <li>RLM-fold2-base-binidx-5-ms-1234-ds-1234.tar.gz</li> <li>rPPB-fold0-base-binidx-5-ms-1234-ds-1234.tar.gz</li> <li>rPPB-fold1-base-binidx-5-ms-1234-ds-1234.tar.gz</li> <li>rPPB-fold2-base-binidx-5-ms-1234-ds-1234.tar.gz</li> <li>SOL-fold0-base-binidx-0-ms-1234-ds-2345.tar.gz</li> <li>SOL-fold0-base-binidx-3-ms-1234-ds-1234.tar.gz</li> <li>SOL-fold0-base-binidx-5-ms-1234-ds-1234.tar.gz</li> <li>SOL-fold1-base-binidx-0-ms-1234-ds-2345.tar.gz</li> <li>SOL-fold1-base-binidx-3-ms-1234-ds-1234.tar.gz</li> <li>SOL-fold1-base-binidx-5-ms-1234-ds-1234.tar.gz</li> <li>SOL-fold2-base-binidx-0-ms-1234-ds-2345.tar.gz</li> <li>SOL-fold2-base-binidx-3-ms-1234-ds-1234.tar.gz</li> <li>SOL-fold2-base-binidx-5-ms-1234-ds-1234.tar.gz</li> </ul> <p><strong>NOTES: </strong> </p> <ul> <li>The hyperparameter search and finetuning scripts are based on Biogen's codebases. As a result, the program design and many software development choices were not based on our own preferences or recommended best practices.</li> <li>As Base-BERT checkpoint files are very large, we have removed them and their corresponding W&B artifact cache files from the compressed files. Users can find the best cross-validated model checkpoints and their re-fitted variants, used for testing, in their corresponding Zenodo records.</li> </ul> <p> </p> </div> <div> <div> <div><strong>Preprint: </strong>https://doi.org/10.48550/arXiv.2603.13627</div> </div> </div>
format Recurso digital
id zenodo_https___doi_org_10_5281_zenodo_19371494
institution Zenodo
language eng
publishDate 2026
publisher Zenodo
record_format zenodo
spellingShingle BERTology of Molecular Property Prediction (Cross-validated Hyperparameter Search: Base-BERT)
Mostafanejad, Mohammad
Saxe, Paul
Crawford, T. Daniel
Molecular property prediction
Masked language modeling
BERT
ADME
Pretraining
Finetuning
Chemical language models
<div> <h1>Cross-validated Hyperparameter Search</h1> <h2>Base-BERT</h2> This record pertains to Cross-validated Hyperparameter Search Experiments involving Base-BERT.</div> <div>It includes the following files:</div> <div> <ul> <li>HLM-fold0-base-binidx-0-ms-1234-ds-2345.tar.gz</li> <li>HLM-fold0-base-binidx-3-ms-1234-ds-1234.tar.gz</li> <li>HLM-fold0-base-binidx-5-ms-1234-ds-1234.tar.gz</li> <li>HLM-fold1-base-binidx-0-ms-1234-ds-2345.tar.gz</li> <li>HLM-fold1-base-binidx-3-ms-1234-ds-1234.tar.gz</li> <li>HLM-fold1-base-binidx-5-ms-1234-ds-1234.tar.gz</li> <li>HLM-fold2-base-binidx-0-ms-1234-ds-2345.tar.gz</li> <li>HLM-fold2-base-binidx-3-ms-1234-ds-1234.tar.gz</li> <li>HLM-fold2-base-binidx-5-ms-1234-ds-1234.tar.gz</li> <li>hPPB-fold0-base-binidx-0-ms-1234-ds-2345.tar.gz</li> <li>hPPB-fold0-base-binidx-3-ms-1234-ds-1234.tar.gz</li> <li>hPPB-fold0-base-binidx-5-ms-1234-ds-1234.tar.gz</li> <li>hPPB-fold1-base-binidx-0-ms-1234-ds-2345.tar.gz<br>hPPB-fold1-base-binidx-3-ms-1234-ds-1234.tar.gz</li> <li>hPPB-fold1-base-binidx-5-ms-1234-ds-1234.tar.gz</li> <li>hPPB-fold2-base-binidx-0-ms-1234-ds-2345.tar.gz</li> <li>hPPB-fold2-base-binidx-3-ms-1234-ds-1234.tar.gz</li> <li>hPPB-fold2-base-binidx-5-ms-1234-ds-1234.tar.gz</li> <li>MDR1_ER-fold0-base-binidx-5-ms-1234-ds-1234.tar.gz</li> <li>MDR1_ER-fold1-base-binidx-5-ms-1234-ds-1234.tar.gz</li> <li>MDR1_ER-fold2-base-binidx-5-ms-1234-ds-1234.tar.gz</li> <li>RLM-fold0-base-binidx-5-ms-1234-ds-1234.tar.gz</li> <li>RLM-fold1-base-binidx-5-ms-1234-ds-1234.tar.gz</li> <li>RLM-fold2-base-binidx-5-ms-1234-ds-1234.tar.gz</li> <li>rPPB-fold0-base-binidx-5-ms-1234-ds-1234.tar.gz</li> <li>rPPB-fold1-base-binidx-5-ms-1234-ds-1234.tar.gz</li> <li>rPPB-fold2-base-binidx-5-ms-1234-ds-1234.tar.gz</li> <li>SOL-fold0-base-binidx-0-ms-1234-ds-2345.tar.gz</li> <li>SOL-fold0-base-binidx-3-ms-1234-ds-1234.tar.gz</li> <li>SOL-fold0-base-binidx-5-ms-1234-ds-1234.tar.gz</li> <li>SOL-fold1-base-binidx-0-ms-1234-ds-2345.tar.gz</li> <li>SOL-fold1-base-binidx-3-ms-1234-ds-1234.tar.gz</li> <li>SOL-fold1-base-binidx-5-ms-1234-ds-1234.tar.gz</li> <li>SOL-fold2-base-binidx-0-ms-1234-ds-2345.tar.gz</li> <li>SOL-fold2-base-binidx-3-ms-1234-ds-1234.tar.gz</li> <li>SOL-fold2-base-binidx-5-ms-1234-ds-1234.tar.gz</li> </ul> <p><strong>NOTES: </strong> </p> <ul> <li>The hyperparameter search and finetuning scripts are based on Biogen's codebases. As a result, the program design and many software development choices were not based on our own preferences or recommended best practices.</li> <li>As Base-BERT checkpoint files are very large, we have removed them and their corresponding W&B artifact cache files from the compressed files. Users can find the best cross-validated model checkpoints and their re-fitted variants, used for testing, in their corresponding Zenodo records.</li> </ul> <p> </p> </div> <div> <div> <div><strong>Preprint: </strong>https://doi.org/10.48550/arXiv.2603.13627</div> </div> </div>
title BERTology of Molecular Property Prediction (Cross-validated Hyperparameter Search: Base-BERT)
topic Molecular property prediction
Masked language modeling
BERT
ADME
Pretraining
Finetuning
Chemical language models
url https://doi.org/10.5281/zenodo.19371494