A general-purpose text embedding model from the GTE family for multilingual semantic similarity tasks.