LLMs

• Generative AI Applications

• Finetuning open source LLMs such as Mistral 7B, Llama 7B, Zephyr 7B

• Integrating RAG and chat memory into the finetuned LLMs

• Performing inference with Llamacpp on local machines including both CPU and GPU

• Employing OpenAI models with finetuning + RAG as personal assistants.

• Deploying LLMs as web apps using React and Flask on Azure Web services.

Use

• In order to perform fine-tuning and the inference, the scripts that are in FineTune/INSTRUCT/ directory can be used.
• RAG performance can be tested using the jupyter notebook RAG/RAG_faiss.ipynb
• The assistant can also be performed using the full-stack application which can be found under Chatbot_Apps/Flask_React_chat directory.