Hugging Face's logo

Datasets:
ko-vlm
/
KVQA-ChatML

Tasks:
Visual Question Answering
Modalities:
Image
Text
Formats:
parquet
Sub-tasks:
visual-question-answering
Languages:
Korean
Size:
10K - 100K
Tags:
korean
vqa
visual-question-answering
multimodal
chatml
conversation
Libraries:
Datasets
Dask
License:
Dataset card Data Studio Files
xet
Community
1
HwangJohn commited on
Commit
cf0ca6b
·
verified ·
1 Parent(s): d19ab5a

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +6 -4
README.md CHANGED
@@ -14,20 +14,22 @@ pretty_name: KVQA (Korean VQA) - ChatML Format
14
  dataset_info:
15
  config_name: default
16
  features:
17
- - name: messages
18
  sequence:
19
- - name: role
20
- dtype: string
21
  - name: content
22
  dtype: string
 
 
23
  - name: image
24
  dtype: image
 
 
25
  - name: source
26
  dtype: string
27
  - name: answer_type
28
  dtype: string
29
  - name: answerable
30
- dtype: bool
31
  num_examples: 100445
32
  num_rows: 100445
33
  download_size: 0
 
14
  dataset_info:
15
  config_name: default
16
  features:
17
+ - name: conversations
18
  sequence:
 
 
19
  - name: content
20
  dtype: string
21
+ - name: role
22
+ dtype: string
23
  - name: image
24
  dtype: image
25
+ - name: id
26
+ dtype: string
27
  - name: source
28
  dtype: string
29
  - name: answer_type
30
  dtype: string
31
  - name: answerable
32
+ dtype: int64
33
  num_examples: 100445
34
  num_rows: 100445
35
  download_size: 0