Molecular Sciences Software Institute
Published March 11, 2026 | Version 03-17-2026
Software Open

BERTology of Molecular Property Prediction (Randomness: Tiny-BERT)

Authors/Creators

  • 1. ROR icon Molecular Sciences Software Institute
  • 2. ROR icon Virginia Tech

Description

Randomness in Model Initialization and Data Sampling Experiments

Tiny-BERT

This record pertains to Randomness in Model Initialization and Data Sampling Experiments.
It involves the following files:
  • berttiny-ms-1234-ds-1234.tar.gz
  • berttiny-ms-1234-ds-2345.tar.gz
  • berttiny-ms-1234-ds-3456.tar.gz
  • berttiny-ms-2345-ds-1234.tar.gz
  • berttiny-ms-3456-ds-1234.tar.gz
Each tar file contains all model artifacts (checkpoints, random-number generator states, optimizer states etc.), training logs (Tensorboard, MLFlow and Weights & Biases), and evaluation results, configuration files, run scripts, SLURM sbatch driver scripts, and any additional artifacts generated during the experiments.
 
Preprint: https://doi.org/10.48550/arXiv.2603.13627

Files

Files (2.6 GB)

Name Size Download all
md5:ddcf158c4ed73231efaa6474552b030f
526.2 MB Download
md5:8a3e78a33c672761311f5c7723dc1e4e
526.9 MB Download
md5:428a09e5a511c50b0b33b55c3445c046
527.0 MB Download
md5:18c257c80784bd534a02c8e6d20110c2
526.6 MB Download
md5:d2ed8ac98efafa7a248920065b0fc3fb
526.2 MB Download

Additional details

Software

Repository URL
https://github.com/molssi-ai/bertology
Programming language
Python
Development Status
Active