This model has been created by keeping only Persian and English token embeddings in mT5-small model. The model was created by following instruction from this link
