File size: 1,406 Bytes
a4e24af
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
e4345b5
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
---
pretty_name: Shaft Process Dataset (5000)
license: other
task_categories:
  - image-to-text
language:
  - en
size_categories:
  - 1K<n<10K
annotations_creators:
  - machine-generated
source_datasets:
  - original
multilinguality:
  - monolingual
configs:
  - config_name: default
    data_files:
      - split: train
        path: train.jsonl
      - split: validation
        path: val.jsonl
      - split: test
        path: test.jsonl
tags:
  - multimodal
  - manufacturing
  - cad
  - cnc
---


# Shaft Dataset (70/30 single vs multi-image)

Tento dataset vznikl automaticky z generovaných výkresů hřídelí.

- **Počet záznamů (po spárování)**: 5000

- **Obrázky**:

  - `images/Shaft_XXXXX.png` — celý výkres (export přes Inkscape)

  - `detail_png/Shaft_XXXXX_length.png` — výřez délkových kót

  - `detail_png/Shaft_XXXXX_diam.png` — výřez průměrových kót

  - `detail_png/Shaft_XXXXX_title.png` — výřez popisového pole

- **Mix**: cca 70 % má 1 obrázek, cca 30 % má 4 obrázky


## Formát JSONL

Každý řádek:

```json
{
  "id": "Shaft_01234",
  "images": ["images/Shaft_01234.png"],
  "conversations": [
    {"from": "user", "value": "<image>\nGenerate the machining process for the shaft in the image. Use the exact 10–11 step template."},
    {"from": "assistant", "value": "1. ...\n2. ..."}
  ]
}
```

## Split
- train: 89%
- val: 5%
- test: 5%