Molecular Datasets
Collection
Small-molecule and polymer datasets. Start from molecular-datasets-index. https://liugangcode.github.io/molecular-dataset-hub/ • 220 items • Updated
Datasets:
You need to agree to share your Hugging Face information to access this dataset
This repository is publicly accessible, but you have to accept the conditions to access its files and content.
Review the dataset and component licenses below before requesting access.
Log in or Sign Up to review the conditions and access this dataset content.
polyVERSE ionization energy
polyVERSE polymer data for ionization energy. Records retain polymer representations, source property names, and available composition or measurement conditions.
Part of the Molecular Datasets collection.
Get started
from datasets import load_dataset
dataset = load_dataset("liuganghuggingface/moldata-polyverse_ionization", "000_data")
Choose a configuration and split from the table below. Target names and units follow the source dataset.
Where the data comes from
- Original data: https://github.com/Ramprasad-Group/polyVERSE/tree/main/Other/Ionization_energy
- Original study: polyVERSE ionization energy
Tables
| Config | Splits | Rows | Targets |
|---|---|---|---|
000_data |
full | 370 | value (regression; units not specified) |
data/ contains rows with readable molecular structures; identical observations are collapsed. observations/ retains all source rows, including unreadable structures. Source label values are preserved without averaging or filling missing values.
Source columns are stored as text to retain values such as <10. Convert numeric labels explicitly and handle censored or missing values for your task. SMILES_original retains the input representation; SMILES_canonical provides a canonical representation. Use structure_key to group molecular components when checking duplicate molecules or overlap between splits. structure_status identifies structure parsing outcomes. Salts, stereochemistry, and polymer attachment points remain relevant to interpretation.
Choosing an evaluation setup
These indicators summarize the available structure and label checks; they are not a guarantee of suitability for every task.
- computed: Measurements, calculations, and model predictions should be evaluated separately.
- valid structures: Rows with unreadable molecular structures remain available in the observations files.
- 370 labeled molecules: Median molecules with a label per task.
Usage notes and limitations
- Task-specific provenance and units must be taken from original column names and adjacent README.
- Repeated structures at different conditions or compositions are not duplicate measurements.
- Repository code license does not supersede the GTRC data terms.
License
GTRC General Public Use License Agreement; noncommercial resale restrictions; retain complete license
Sign in and accept the applicable source terms to request access. Each source or component keeps its own license.
Citation
Cite the original study and data source:
- Downloads last month
- 26