Downloads · 30 days
20
1% of all-time downloads
lachkarsalim/Helsinki-translation-English_Moroccan-Arabic
Helsinki-translation-English_Moroccan-Arabic is a machine learning model from lachkarsalim. Use it for the machine learning task on the model card, and read the license before you ship it in a product. It is set up for transformers.
Model Description : This model is a fine-tuned version of a Transformer Model Helsinki-en-ar, adapted to translate text from English to Darija (Moroccan Arabic) . The fine-tuning process was conducted on a substantial…
Downloads · 30 days
20
1% of all-time downloads
All-time downloads
2K
Public
Parameters
239M
2.9 GB on disk
Likes
9
Public
Click a slice to open those files.
.pt1.9 GB · 67%
From the Hugging Face model README
Model Description : This model is a fine-tuned version of a Transformer Model Helsinki-en-ar, adapted to translate text from English to Darija (Moroccan Arabic) . The fine-tuning process was conducted on a substantial dataset.
Fine-tuning Details Source Model: Helsinki-NLP/opus-mt-en-ar Fine-tuning Objective: To adapt the pre-existing English to Arabic translation model to perform translations from English to Arabic Darija . Dataset: The model was fine-tuned using the Darija Open Dataset (DODa), an open-source project dedicated to the Moroccan dialect. DODa contains approximately 150,000 entries, making it one of the largest open-source collaborative projects for Darija <=> English translation aimed at Natural Language Processing (NLP) applications. Training Examples: More than 15,000 translation pairs from Darija to Arabic were used for fine-tuning. Training Time: The fine-tuning process took approximately 8 hours to complete
Acknowledgments : I would like to acknowledge the contributors to the Darija Open Dataset (DODa) for providing an extensive and valuable resource for training this model. Their effort in building the largest open-source Darija dataset has significantly facilitated research and development in NLP applications tailored to Moroccan Arabic.