Datasets:
Release 2026-08-19 with A/B/C evaluation subsets
Browse files- README.md +9 -7
- data/test-00000-of-00004.parquet +2 -2
- data/test-00001-of-00004.parquet +2 -2
- data/test-00002-of-00004.parquet +2 -2
- data/test-00003-of-00004.parquet +2 -2
- release_manifest.json +11 -11
README.md
CHANGED
|
@@ -64,13 +64,13 @@ dataset_info:
|
|
| 64 |
list: float64
|
| 65 |
---
|
| 66 |
|
| 67 |
-
# MolRecBench-Wild (`
|
| 68 |
|
| 69 |
MolRecBench-Wild is a real-world benchmark for optical chemical structure
|
| 70 |
recognition. This release contains **5,024** molecular structure images and
|
| 71 |
their CARBON molecular-graph annotations from **818** source articles.
|
| 72 |
|
| 73 |
-
This is the repository's authoritative `
|
| 74 |
5,029-sample snapshot described in the first arXiv version of the paper.
|
| 75 |
|
| 76 |
## Dataset structure
|
|
@@ -92,17 +92,19 @@ print(sample["id"], sample["symbols"], sample["bonds"])
|
|
| 92 |
|
| 93 |
- `image`: cropped molecular structure image.
|
| 94 |
- `id`: stable sample identifier and original image filename.
|
| 95 |
-
- `release_id`: dataset release
|
| 96 |
- `source`, `source_doi`, `source_url`: provenance derived from the DOI encoded
|
| 97 |
in the sample ID.
|
| 98 |
-
- `evaluation_subset`: benchmark subset (`
|
| 99 |
-
`no_property_gt2`, or `property`).
|
| 100 |
- `hardcase_label`: visual and chemical difficulty labels.
|
| 101 |
- `symbols`, `charges`, `radicals`, `valences`, `isotopes`, `attach_points`,
|
| 102 |
`coords`, `bonds`, `brackets`: CARBON graph annotation fields.
|
| 103 |
|
| 104 |
-
Subset sizes are 1,987 `
|
| 105 |
-
|
|
|
|
|
|
|
|
|
|
| 106 |
|
| 107 |
## Source information
|
| 108 |
|
|
|
|
| 64 |
list: float64
|
| 65 |
---
|
| 66 |
|
| 67 |
+
# MolRecBench-Wild (`2026-08-19`)
|
| 68 |
|
| 69 |
MolRecBench-Wild is a real-world benchmark for optical chemical structure
|
| 70 |
recognition. This release contains **5,024** molecular structure images and
|
| 71 |
their CARBON molecular-graph annotations from **818** source articles.
|
| 72 |
|
| 73 |
+
This is the repository's authoritative `2026-08-19` release. It differs from the
|
| 74 |
5,029-sample snapshot described in the first arXiv version of the paper.
|
| 75 |
|
| 76 |
## Dataset structure
|
|
|
|
| 92 |
|
| 93 |
- `image`: cropped molecular structure image.
|
| 94 |
- `id`: stable sample identifier and original image filename.
|
| 95 |
+
- `release_id`: dataset release date (`2026-08-19`, ISO 8601 format).
|
| 96 |
- `source`, `source_doi`, `source_url`: provenance derived from the DOI encoded
|
| 97 |
in the sample ID.
|
| 98 |
+
- `evaluation_subset`: benchmark difficulty subset (`A`, `B`, or `C`).
|
|
|
|
| 99 |
- `hardcase_label`: visual and chemical difficulty labels.
|
| 100 |
- `symbols`, `charges`, `radicals`, `valences`, `isotopes`, `attach_points`,
|
| 101 |
`coords`, `bonds`, `brackets`: CARBON graph annotation fields.
|
| 102 |
|
| 103 |
+
Subset sizes are 1,987 `A` samples, 1,976 `B` samples, and 1,061 `C` samples.
|
| 104 |
+
Subset A has no specified chemical-semantic property difficulty labels and
|
| 105 |
+
relatively few visual difficulty labels; B has no such property labels but
|
| 106 |
+
more visual difficulty labels; C contains specified chemical-semantic property
|
| 107 |
+
difficulty labels.
|
| 108 |
|
| 109 |
## Source information
|
| 110 |
|
data/test-00000-of-00004.parquet
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:94ce9705fc330fca6512ab5fa3aff6ba9eceedacfd0cce607254cb6a5ea9332a
|
| 3 |
+
size 21383248
|
data/test-00001-of-00004.parquet
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:74d177f6470d63b87a5f729ba0b0c59ff997bfcd5fedb033ab6f3b14e4b4a433
|
| 3 |
+
size 22626400
|
data/test-00002-of-00004.parquet
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9858f4bab5fe35d691ea50b2170530e7be8507220f04df422dcad1d2735125b3
|
| 3 |
+
size 23655556
|
data/test-00003-of-00004.parquet
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:78131a479bfc62006d5ba252081227d30c59d830b708b61595968a9791c82788
|
| 3 |
+
size 20795731
|
release_manifest.json
CHANGED
|
@@ -1,42 +1,42 @@
|
|
| 1 |
{
|
| 2 |
"schema_version": 1,
|
| 3 |
-
"release_id": "
|
| 4 |
"sample_count": 5024,
|
| 5 |
"source_doi_count": 818,
|
| 6 |
-
"annotation_jsonl_sha256": "
|
| 7 |
-
"image_set_sha256": "
|
| 8 |
"parquet_shards": [
|
| 9 |
{
|
| 10 |
"path": "data/test-00000-of-00004.parquet",
|
| 11 |
"rows": 1256,
|
| 12 |
"row_groups": 13,
|
| 13 |
"max_rows_per_group": 100,
|
| 14 |
-
"size_bytes":
|
| 15 |
-
"sha256": "
|
| 16 |
},
|
| 17 |
{
|
| 18 |
"path": "data/test-00001-of-00004.parquet",
|
| 19 |
"rows": 1256,
|
| 20 |
"row_groups": 13,
|
| 21 |
"max_rows_per_group": 100,
|
| 22 |
-
"size_bytes":
|
| 23 |
-
"sha256": "
|
| 24 |
},
|
| 25 |
{
|
| 26 |
"path": "data/test-00002-of-00004.parquet",
|
| 27 |
"rows": 1256,
|
| 28 |
"row_groups": 13,
|
| 29 |
"max_rows_per_group": 100,
|
| 30 |
-
"size_bytes":
|
| 31 |
-
"sha256": "
|
| 32 |
},
|
| 33 |
{
|
| 34 |
"path": "data/test-00003-of-00004.parquet",
|
| 35 |
"rows": 1256,
|
| 36 |
"row_groups": 13,
|
| 37 |
"max_rows_per_group": 100,
|
| 38 |
-
"size_bytes":
|
| 39 |
-
"sha256": "
|
| 40 |
}
|
| 41 |
]
|
| 42 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"schema_version": 1,
|
| 3 |
+
"release_id": "2026-08-19",
|
| 4 |
"sample_count": 5024,
|
| 5 |
"source_doi_count": 818,
|
| 6 |
+
"annotation_jsonl_sha256": "9b6f37a1b9c278b5dedf1e78ddaf9505cc707fd99f23976b27acf0fbe5d38958",
|
| 7 |
+
"image_set_sha256": "90322952a53406d381c9bf4e852c96f76fffe98c04b43c0d3a60431aea90224f",
|
| 8 |
"parquet_shards": [
|
| 9 |
{
|
| 10 |
"path": "data/test-00000-of-00004.parquet",
|
| 11 |
"rows": 1256,
|
| 12 |
"row_groups": 13,
|
| 13 |
"max_rows_per_group": 100,
|
| 14 |
+
"size_bytes": 21383248,
|
| 15 |
+
"sha256": "94ce9705fc330fca6512ab5fa3aff6ba9eceedacfd0cce607254cb6a5ea9332a"
|
| 16 |
},
|
| 17 |
{
|
| 18 |
"path": "data/test-00001-of-00004.parquet",
|
| 19 |
"rows": 1256,
|
| 20 |
"row_groups": 13,
|
| 21 |
"max_rows_per_group": 100,
|
| 22 |
+
"size_bytes": 22626400,
|
| 23 |
+
"sha256": "74d177f6470d63b87a5f729ba0b0c59ff997bfcd5fedb033ab6f3b14e4b4a433"
|
| 24 |
},
|
| 25 |
{
|
| 26 |
"path": "data/test-00002-of-00004.parquet",
|
| 27 |
"rows": 1256,
|
| 28 |
"row_groups": 13,
|
| 29 |
"max_rows_per_group": 100,
|
| 30 |
+
"size_bytes": 23655556,
|
| 31 |
+
"sha256": "9858f4bab5fe35d691ea50b2170530e7be8507220f04df422dcad1d2735125b3"
|
| 32 |
},
|
| 33 |
{
|
| 34 |
"path": "data/test-00003-of-00004.parquet",
|
| 35 |
"rows": 1256,
|
| 36 |
"row_groups": 13,
|
| 37 |
"max_rows_per_group": 100,
|
| 38 |
+
"size_bytes": 20795731,
|
| 39 |
+
"sha256": "78131a479bfc62006d5ba252081227d30c59d830b708b61595968a9791c82788"
|
| 40 |
}
|
| 41 |
]
|
| 42 |
}
|