davanstrien HF Staff commited on
Commit
161c45e
·
verified ·
1 Parent(s): e73b02e

Upload README.md with huggingface_hub

Browse files
Files changed (1) hide show
  1. README.md +58 -23
README.md CHANGED
@@ -1,25 +1,60 @@
1
  ---
2
- dataset_info:
3
- features:
4
- - name: image
5
- dtype: image
6
- - name: source_file
7
- dtype: string
8
- - name: card_id
9
- dtype: string
10
- - name: extraction
11
- dtype: string
12
- - name: inference_info
13
- dtype: string
14
- splits:
15
- - name: train
16
- num_bytes: 81241528
17
- num_examples: 120
18
- download_size: 81194148
19
- dataset_size: 81241528
20
- configs:
21
- - config_name: default
22
- data_files:
23
- - split: train
24
- path: data/train-*
25
  ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
  ---
2
+ tags:
3
+ - ocr
4
+ - structured-extraction
5
+ - document-processing
6
+ - nuextract3
7
+ - markdown
8
+ - uv-script
9
+ - generated
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
10
  ---
11
+
12
+ # NuExtract3 on davanstrien/bpl-shelf-list-sample
13
+
14
+ This dataset contains outputs from [davanstrien/bpl-shelf-list-sample](https://huggingface.co/datasets/davanstrien/bpl-shelf-list-sample) processed with [NuExtract3](https://huggingface.co/numind/NuExtract3), a 4B vision-language model for document understanding.
15
+
16
+ ## Processing Details
17
+
18
+ - **Source Dataset**: [davanstrien/bpl-shelf-list-sample](https://huggingface.co/datasets/davanstrien/bpl-shelf-list-sample)
19
+ - **Model**: [numind/NuExtract3](https://huggingface.co/numind/NuExtract3)
20
+ - **Mode**: structured-extraction
21
+ - **Number of Samples**: 120
22
+ - **Processing Time**: 5.1 min
23
+ - **Processing Date**: 2026-05-21 10:05 UTC
24
+
25
+ ### Configuration
26
+
27
+ - **Image Column**: `image`
28
+ - **Output Column**: `extraction`
29
+ - **Dataset Split**: `train`
30
+ - **Temperature**: 0.2
31
+ - **Thinking Mode**: disabled
32
+
33
+ ### Extraction Template
34
+
35
+ ```json
36
+ {
37
+ "card_type": [
38
+ "bibliographic",
39
+ "shelf_divider",
40
+ "other"
41
+ ],
42
+ "shelf_no": "verbatim-string",
43
+ "volumes": "string",
44
+ "additions": "string",
45
+ "author": "verbatim-string",
46
+ "title": "verbatim-string",
47
+ "place_of_publication": "verbatim-string",
48
+ "date": "string",
49
+ "accession_no": "verbatim-string"
50
+ }
51
+ ```
52
+
53
+ ## Dataset Structure
54
+
55
+ Original columns plus:
56
+ - `extraction`: NuExtract3 output (JSON string)
57
+ - `inference_info`: JSON list tracking models applied to this dataset
58
+
59
+
60
+ Generated with [UV Scripts](https://huggingface.co/uv-scripts)