# Datasets

:::{margin}
```{image} ../../assets/icons/catalysis.svg
:alt: Catalysis
:width: 100px
```
:::

This section provides documentation for all catalyst-related datasets in the Open Catalyst Project. These datasets are used to train and evaluate machine learning models for heterogeneous catalysis applications.

:::{tip}
For most new users, we recommend starting with the [UMA model](../../core/uma), which has been trained on all of these datasets and provides state-of-the-art performance.
:::

::::{grid} 2 2 3 3

:::{grid-item-card} OC20
:link: oc20

The foundational Open Catalyst 2020 dataset with 133M+ DFT calculations for adsorbate-surface systems.
:::

:::{grid-item-card} OC20-mAds
:link: oc20_mads

Multi-adsorbate extension of OC20 including coverage effects on catalyst surfaces.
:::

:::{grid-item-card} OC20Dense
:link: oc20dense

Dense sampling of adsorbate configurations for adsorption energy calculations.
:::

:::{grid-item-card} OC20NEB
:link: oc20neb

NEB trajectories for transition state calculations including desorptions, dissociations, and transfers.
:::

:::{grid-item-card} OC22
:link: oc22

Open Catalyst 2022 dataset focusing on oxide electrocatalysts with total energy predictions.
:::

:::{grid-item-card} OC25
:link: oc25

Solid-liquid interface dataset with 8M DFT calculations for electrocatalysis applications.
:::

:::{grid-item-card} OCx24
:link: ocx24

Experimental validation dataset bridging computational and experimental catalysis research.
:::

::::