Datasets:
Modalities:
Text
Formats:
parquet
Sub-tasks:
semantic-similarity-classification
Size:
1M - 10M
Tags:
paraphrase-generation
License:
Which model to use with the Tapaco dataset in French?
#1 opened almost 2 years ago
by
LeMoussel
