RAG - Retrieval-Augmented Generation Application

• Persionalized RAG for query retrievals.
• Quick Q/A with the given knowledge bases.

---

Demo Images:

Tech stack

Libraries used to get done the project:

PDFLoader - to purse the data
TextSplit(RecursiveCharacterTextSplitter) - to split the text into chunks
sentence-transformers(all-mpnet-base-v2) - Embedding the chunks
ChromaDB - to store the Embedding Vectors
Groq-LLM(openai/gpt-oss-120b) - to summarize the context along with the prompt

---

Workflow of the RAG Pipeline

Project setup Locally:

Go to Terminal or bash of you project folder:

git clone https://github.com/praveensunkara19/RAG.git

cd RAG

python -m venv myenv 

myenv/scripts/activate

pip install -r requirements.txt

streamlit run app.py

Future scope:

• Adding persionlization like speech
• Retrieval of images - videos
• Industry level optimization for better retrieval with - FAISS, PineCone
• Works with all kind of document types.

--------------------- "References" ----------------

1. https://huggingface.co/MBZUAI/LaMini-T5-738M LLM to process the chunks of the vectorised db to give the results
2. https://www.trychroma.com/ for chromadb and related libraries.
3. https://python.langchain.com/v0.2/docs/integrations/platforms/huggingface/ for HuggingFace LLMs and Endpoints
4. https://pypi.org/project/sentence-transformers/ for the SentenceTransformerEmbeddings used as the text_spitting
5. https://docs.streamlit.io/ used to display the output (UI)