Optimal Train/Test Split
TAL-SCQ5K
151
•
341
Apache-2.0
10K – 100K
Image-to-Text
Image Classification
English
Chinese
Japanese
Polish
Opened about a year ago
by @ceivin-7916
0
@ceivin-7916a year ago
The balanced 3K/2K train/test split makes this dataset perfect for both robust training and systematic evaluation!