Modelpublic

CATIE-AQ/french_paraphrase_flan-t5-large

sourceHugging Faceupdated 10mo agoView on Hugging Face

2likes27downloads

Model Card

Description

Google's flan-t5-large model finetuned on frenchPARAPHRASE) dataset to generate paraphrases for French texts.

Scores on the test split of frenchPARAPHRASE: | Model | ROUGE-Lsum | ROUGE-L | ROUGE-2 | ROUGE-1 | ParaScore | | :--- | :---: | :---: | :---: | :---: | :---: | | CATIE-AQ/french_paraphrase_flan-t5-base | 0.5772 | 0.5771 | 0.4423 | 0.6069 | 0.7561 | | CATIE-AQ/french_paraphrase_flan-t5-large | 0.5887 | 0.5887 | 0.4752 | 0.6160 | 0.7590 |

Usage

from transformers import AutoTokenizer, AutoModelForSeq2SeqLM

tokenizer = AutoTokenizer.from_pretrained("CATIE-AQ/french_paraphrase_flan-t5-large")
model = AutoModelForSeq2SeqLM.from_pretrained("CATIE-AQ/french_paraphrase_flan-t5-large")