# Context-aware translation

TowerInstruct-13B is a 13-billion-parameter language model fine-tuned from the TowerBase model on the TowerBlocks supervised fine-tuning dataset, specifically designed for handling various translation-related tasks.

TowerInstruct-7B-v0.2 is a 7-billion-parameter multilingual large language model focused on translation-related tasks, supporting 10 languages.

This model is based on Helsinki-NLP/opus-mt-en-fr, further trained for English-French translation with context tag format on the IWSLT17 dataset, and fine-tuned on the SCAT+ training set.

This model is based on Helsinki-NLP/opus-mt-en-fr, further trained on the IWSLT17 dataset using context tag format for English-French translation, and fine-tuned on the SCAT+ dataset. Suitable for both context-aware and context-free English-French translation tasks.