Datasets:
Tasks:
Text Classification
Modalities:
Text
Formats:
parquet
Sub-tasks:
multi-input-text-classification
Languages:
Catalan
Size:
10K - 100K
License:
data/train-00000-of-00001.parquet filter=lfs diff=lfs merge=lfs -text | |