ChemFuse

# ChemFuse Datasets This repository contains the datasets used for training, evaluation, and downstream tasks in the ChemFuse. ## Unlabelled Polymer Dataset The [polymer dataset](https://huggingface.co/datasets/jablonkagroup/ChemFuse/tree/main/polymer_data) contains unlabelled polymer entries with three different identifiers: PSMILES, BigSMILES, and polymer names. ## Unlabelled MoF Dataset The [MOF dataset](https://huggingface.co/datasets/jablonkagroup/ChemFuse/tree/main/MOF_data) contains different modalities---MOFid, pXRD, and crystal structures — in a specific directory, with indices in the same order. ## Downstream Property Datasets Each downstream dataset includes property labels along with the corresponding learned embeddings. ### Polymers - [**Glass transition temperature**](https://huggingface.co/datasets/jablonkagroup/ChemFuse/tree/main/polymer_glass_transition_temperature) — with PSMILES and BigSMILES embeddings ### Molecules - [**Melting point**](https://huggingface.co/datasets/jablonkagroup/ChemFuse/tree/main/molecule_melting_point) — with SMILES and HNMR embeddings - [**Solubility**](https://huggingface.co/datasets/jablonkagroup/ChemFuse/tree/main/solubility) ([AqSolDB](https://huggingface.co/datasets/jablonkagroup/ChemFuse/tree/main/solubility_aquasoldb)) — with SMILES and HNMR embeddings - [**Quantitative estimation of drug-likeness**](https://huggingface.co/datasets/jablonkagroup/ChemFuse/tree/main/quantitative_estimation_drug_likeliness) — with SMILES and HNMR embeddings ### Metal–Organic Frameworks (MOFs) - [**Pure_methane_uptake_capacity**](https://huggingface.co/datasets/jablonkagroup/ChemFuse/tree/main/pure_uptake_methane_298.00_6500000) — with MOFid and pXRD embeddings - [**Pure carbondioxide uptake_capacity**](https://huggingface.co/datasets/jablonkagroup/ChemFuse/tree/main/pure_uptake_CO2_298.00_1600000) — with MOFid and pXRD embeddings ## Citation If you use these datasets in your work, please cite: ```bibtex } ```