NOTE: This project is in progress.
vi-gemma-2-2b-RAG is a large language model refined from the google/gemma-2-2b-it base model using the LoRA technique. The model is trained on a Vietnamese dataset to improve its ability to process the Vietnamese language and enhance performance in open information retrieval tasks (Retrieval Augmented Generation - RAG).
The vi-gemma-2b-RAG model is suitable for the following tasks:
- Question answering based on Vietnamese context.
- Vietnamese text summarization.
- Vietnamese machine translation.
- Other Vietnamese text generation tasks.