File size: 5,157 Bytes
d0a5d19
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
---
license: apache-2.0
tags:
  - OC20
  - Open Catalyst
  - catalysis
  - S2EF
  - UMA
  - extxyz
  - ASE-lmdb
language:
  - en
  - zh
---

<p align="center">
  <strong>
    <span style="font-size: 30px;">OC20</span>
  </strong>
</p>

## Dataset Description

The OC20 dataset is an extxyz subset of the Open Catalyst 2020 (OC20) S2EF (Structure to Energy and Forces) dataset. It contains atomic coordinates, unit-cell information, energies, and forces for adsorption configurations on catalytic material surfaces. Its energy and force fields are consistent with those in the UMA OC20 configuration, making it a standard benchmark dataset for developing and validating machine-learned interatomic potential methods for catalytic materials.

This repository currently contains two splits:

- `s2ef_200k_uncompressed`: Training subset with 2 shards and 10,000 frames in total, with 8–221 atoms per frame.
- `s2ef_val_id_uncompressed`: In-distribution validation subset with 4 shards and 20,000 frames in total, with 8–223 atoms per frame.

The original data comes from the [Open Catalyst Project / fair-chem OC20](https://fair-chem.github.io/oc20/). This repository provides the accompanying data package for the OneScience UMA fine-tuning example.

The current data package contains approximately 8 files totaling approximately 1.2 GB.

## Supported Tasks

This standardized data repository organizes the complete OC20 S2EF subset package, including training and validation sets, to support:
- Fine-tuning machine-learned interatomic potentials on adsorption configurations at catalytic material surfaces
- Training and validation of energy and force prediction models
- Functional testing and benchmark evaluation of the UMA framework

## Dataset Format and Structure

The data files are located in the `data/oc20/` directory and organized into shards by split:

| File Path | Format | Shape / Content | Description |
|---|---|---|---|
| `oc20/s2ef_200k_uncompressed/0.extxyz` | extxyz | `[N_frames, N_atoms]` | Training subset shard 0, 5,000 frames |
| `oc20/s2ef_200k_uncompressed/1.extxyz` | extxyz | `[N_frames, N_atoms]` | Training subset shard 1, 5,000 frames |
| `oc20/s2ef_val_id_uncompressed/0.extxyz` | extxyz | `[N_frames, N_atoms]` | Validation subset shard 0, 5,000 frames |
| `oc20/s2ef_val_id_uncompressed/1.extxyz` | extxyz | `[N_frames, N_atoms]` | Validation subset shard 1, 5,000 frames |
| `oc20/s2ef_val_id_uncompressed/2.extxyz` | extxyz | `[N_frames, N_atoms]` | Validation subset shard 2, 5,000 frames |
| `oc20/s2ef_val_id_uncompressed/3.extxyz` | extxyz | `[N_frames, N_atoms]` | Validation subset shard 3, 5,000 frames |
| `metadata/schema.yaml` | YAML | Data structure description | extxyz field schema definition |
| `metadata/sha256_manifest.txt` | SHA256 | Checksum manifest | File integrity verification |

### extxyz Data Format

Each configuration consists of an atom-count line, a comment line, and atom lines. The comment line contains `Lattice`, `Properties`, `energy`, `free_energy`, and `pbc`. Each atom line contains the element, coordinates, `move_mask`, `tags`, and `forces`.

| Field | Type | Shape | Description |
|---|---|---|---|
| `species` | str | `[N_atoms]` | Element symbols spanning various metals and nonmetals |
| `pos` | float | `[N_atoms, 3]` | Atomic coordinates in Å |
| `move_mask` | bool | `[N_atoms]` | Whether each atom can move, `T` or `F` |
| `tags` | int | `[N_atoms]` | Atomic labels distinguishing adsorbates, surfaces, subsurfaces, and other categories |
| `forces` | float | `[N_atoms, 3]` | Atomic forces in eV/Å |
| `energy` | float | `[1]` | Total configuration energy in eV |
| `free_energy` | float | `[1]` | Configuration free energy in eV |
| `pbc` | bool | `[3]` | Periodic boundary conditions |
| `Lattice` | float | `[3, 3]` | Periodic cell matrix |


## How to Use the Dataset

### Download the Dataset

```bash
hf download --dataset OneScience-Sugon/oc20 --local-dir ./data
```

## Official OneScience Information

| Platform | OneScience Main Repository | Skills Repository |
|---|---|---|
| Gitee | https://gitee.com/onescience-ai/onescience | https://gitee.com/onescience-ai/oneskills |
| GitHub | https://github.com/onescience-ai/OneScience | https://github.com/onescience-ai/oneskills |

## Limitations and License

This dataset is an official S2EF subset of Open Catalyst 2020 (OC20). No new samples were generated, and the original data content was not altered.

- **Original source**: [Open Catalyst Project / fair-chem OC20](https://fair-chem.github.io/oc20/)
- **License**: Apache License 2.0
- **Usage restrictions**: Refer to the original OC20 data source and the information provided by the OneScience repository and ModelScope page

When using the OC20 dataset, cite the original Open Catalyst paper:

- Chanussot, L.; Das, A.; Goyal, S.; Lavril, T.; Shuaibi, M.; Riviere, M.; Tran, K.; Heras-Domingo, J.; Ho, C.; Hu, W.; Palizhati, A.; Sriram, A.; Wood, B.; Yoon, J.; Parikh, D.; Zitnick, C. L.; Ulissi, Z. Open Catalyst 2020 (OC20) Dataset and Community Challenges. *ACS Catal.* **2021**, *11* (10), 6059–6072. https://doi.org/10.1021/acscatal.0c04525