Hugging Face's logo

Datasets:
liuganghuggingface
/
moldata-simpoly_polydata

Tasks:
Tabular Regression
Modalities:
Tabular
Text
Formats:
parquet
Size:
100K - 1M
ArXiv:
Tags:
chemistry
molecules
polymers
simulation
2025
dft
Libraries:
Datasets
Dask
Polars
License:
Dataset card Data Studio Files
xet
 Community
Dataset Viewer
Auto-converted to Parquet Duplicate
poly_id
stringclasses
130 values
offset_dim
stringclasses
4 values
offset
float64
0
6.9
⌀
idx
int64
0
206k
CR
y
0
0
CR
yz
0
1
CR
z
0
2
CR
xyz
0.1
3
CR
y
0.1
4
CR
yz
0.1
5
CR
xyz
0
6
CR
xyz
0.2
7
CR
yz
0.2
8
CR
z
0.1
9
CR
y
0.2
10
CR
z
0.2
11
CR
xyz
0.3
12
CR
y
0.3
13
CR
z
0.3
14
CR
yz
0.3
15
CR
xyz
0.4
16
CR
y
0.4
17
CR
yz
0.5
18
CR
yz
0.4
19
CR
z
0.4
20
CR
y
0.5
21
CR
xyz
0.5
22
CR
z
0.5
23
CR
y
0.6
24
CR
yz
0.6
25
CR
z
0.6
26
CR
xyz
0.6
27
CR
xyz
0.7
28
CR
y
0.7
29
CR
yz
0.7
30
CR
z
0.7
31
CR
xyz
0.8
32
CR
y
0.8
33
CR
z
0.8
34
CR
yz
0.8
35
CR
y
0.9
36
CR
xyz
1
37
CR
yz
0.9
38
CR
z
0.9
39
CR
y
1
40
CR
yz
1
41
CR
xyz
0.9
42
CR
z
1
43
CR
xyz
1.1
44
CR
yz
1.1
45
CR
y
1.1
46
CR
z
1.1
47
CR
xyz
1.2
48
CR
y
1.2
49
CR
yz
1.2
50
CR
xyz
1.3
51
CR
z
1.2
52
CR
y
1.3
53
CR
yz
1.3
54
CR
z
1.3
55
CR
y
1.4
56
CR
yz
1.4
57
CR
z
1.4
58
CR
xyz
1.4
59
CR
y
1.5
60
CR
yz
1.5
61
CR
xyz
1.5
62
CR
y
1.6
63
CR
xyz
1.6
64
CR
yz
1.6
65
CR
z
1.6
66
CR
xyz
1.7
67
CR
z
1.5
68
CR
y
1.7
69
CR
yz
1.7
70
CR
z
1.7
71
CR
yz
1.8
72
CR
xyz
1.8
73
CR
y
1.8
74
CR
z
1.8
75
CR
xyz
1.9
76
CR
y
1.9
77
CR
yz
1.9
78
CR
xyz
2
79
CR
z
1.9
80
CR
y
2
81
CR
z
2
82
CR
xyz
2.1
83
CR
yz
2
84
CR
z
2.1
85
CR
yz
2.1
86
CR
y
2.1
87
CR
xyz
2.2
88
CR
yz
2.2
89
CR
y
2.2
90
CR
z
2.2
91
CR
xyz
2.3
92
CR
yz
2.3
93
CR
y
2.3
94
CR
z
2.3
95
CR
xyz
2.4
96
CR
y
2.4
97
CR
yz
2.4
98
CR
z
2.4
99
End of preview. Expand in Data Studio

SimPoly PolyData: PolyPack and PolyDiss

Microsoft Research native LMDB quantum-chemical polymer datasets, plus compact Parquet metadata tables. PolyPack covers packed polymers; PolyDiss covers dissociation configurations.

Processed copy of simpoly_polydata from the molecular datasets collection. Structures are canonicalized and validated; labels are never averaged or filled. Prefer the original source for citation.

Where the data comes from

Data quality

Grade: unmeasured (rough; rules on the hub page).

  • info: computed. Label origin from the source's tags.
  • warn: multi-source. Compiled from several labs or publications; assay protocols can differ (Walters 2023: 45% of same-compound IC50 pairs from two papers differ by >0.3 log units).

Modalities

Besides SMILES and labels:

Modality Where Content
source files files/README.md (0 MB) Original file, unchanged. No SMILES table was built from it.
source files files/poly_diss.lmdb (225 MB) Original file, unchanged. No SMILES table was built from it.
source files files/poly_diss.parquet (0 MB) Original file, unchanged. No SMILES table was built from it.
source files files/poly_pack.lmdb (1,906 MB) Original file, unchanged. No SMILES table was built from it.
source files files/poly_pack.parquet (1 MB) Original file, unchanged. No SMILES table was built from it.

Files are copied unchanged from the original source:

from huggingface_hub import hf_hub_download
path = hf_hub_download("liuganghuggingface/moldata-simpoly_polydata", "files/README.md", repo_type="dataset")

Known issues

  • Native atomistic geometry, energies and forces must remain in LMDB; Parquet tables are metadata, not a replacement.
  • Training/evaluation partition is defined by source metadata and original code.
  • Do not combine with OPoly26 without auditing possible chemistry overlap and DFT settings.

License

MIT (official model/data repository)

Citation

Cite the original study and data source:
Downloads last month
38
Number of rows:
241,847
Total file size:
2.13 GB

Collection including liuganghuggingface/moldata-simpoly_polydata

Paper for liuganghuggingface/moldata-simpoly_polydata