Build RAG Chatbot with Haystack, Pgvector, STACKIT Meta-Llama-3.1-70B-Instruct-FP8, and Optimum all-mpnet-base-v2

Introduction to RAG

Retrieval-Augmented Generation (RAG) is a game-changer for GenAI applications, especially in conversational AI. It combines the power of pre-trained large language models (LLMs) like OpenAI’s GPT with external knowledge sources stored in vector databases such as Milvus and Zilliz Cloud, allowing for more accurate, contextually relevant, and up-to-date response generation. A RAG pipeline usually consists of four basic components: a vector database, an embedding model, an LLM, and a framework.

Key Components We'll Use for This RAG Chatbot

This tutorial shows you how to build a simple RAG chatbot in Python using the following components:

Haystack: An open-source Python framework designed for building production-ready NLP applications, particularly question answering and semantic search systems. Haystack excels at retrieving information from large document collections through its modular architecture that combines retrieval and reader components. Ideal for developers creating search applications, chatbots, and knowledge management systems that require efficient document processing and accurate information extraction from unstructured text.
Pgvector: an open-source extension for PostgreSQL that enables efficient storage and querying of high-dimensional vector data, essential for machine learning and AI applications. Designed to handle embeddings, it supports fast approximate nearest neighbor (ANN) searches using algorithms like HNSW and IVFFlat. Since it is just a vector search add-on to traditional search rather than a purpose-built vector database, it lacks scalability and availability and many other advanced features required by enterprise-level applications. Therefore, if you prefer a much more scalable solution or hate to manage your own infrastructure, we recommend using Zilliz Cloud, which is a fully managed vector database service built on the open-source Milvus and offers a free tier supporting up to 1 million vectors.)
STACKIT Meta-Llama-3.1-70B-Instruct-FP8: A 70-billion-parameter instruction-tuned AI model optimized for complex NLP tasks. Leveraging FP8 quantization, it balances high performance with efficient resource usage, excelling in understanding nuanced instructions and generating context-aware responses. Ideal for enterprise applications requiring precise, scalable solutions in chatbots, virtual assistants, data analysis, and automated content generation.
Optimum all-mpnet-base-v2: A high-performance sentence-transformers model optimized for semantic textual similarity, offering robust multilingual embeddings. Its strengths include efficient inference, scalability, and state-of-the-art accuracy in tasks like semantic search, clustering, and retrieval-augmented generation (RAG). Ideal for enterprise applications requiring fast, precise text analysis across diverse languages and domains.

By the end of this tutorial, you’ll have a functional chatbot capable of answering questions based on a custom knowledge base.

Note: Since we may use proprietary models in our tutorials, make sure you have the required API key beforehand.

Step 1: Install and Set Up Haystack

import os
import requests

from haystack import Pipeline
from haystack.components.converters import MarkdownToDocument

from haystack.components.preprocessors import DocumentSplitter
from haystack.components.writers import DocumentWriter

Step 2: Install and Set Up STACKIT Meta-Llama-3.1-70B-Instruct-FP8

STACKIT is the cloud and colocation provider of the Schwarz Group. We can use different models on its cloud services with ease through its API.

pip install stackit-haystack

from haystack_integrations.components.generators.stackit import STACKITChatGenerator
from haystack.dataclasses import ChatMessage

generator = STACKITChatGenerator(model="neuralmagic/Meta-Llama-3.1-70B-Instruct-FP8")

Step 3: Install and Set Up Optimum all-mpnet-base-v2

Haystack's OptimumTextEmbedder embeds text strings using models loaded with the HuggingFace Optimum library. It uses the ONNX runtime for high-speed inference. Similarly to other Embedders, this component allows adding prefixes (and suffixes) to include instructions. For more details, refer to the Optimum API Reference.

pip install optimum-haystack

from haystack_integrations.components.embedders.optimum import OptimumTextEmbedder
from haystack.dataclasses import Document
from haystack_integrations.components.embedders.optimum import OptimumDocumentEmbedder

text_embedder = OptimumTextEmbedder(model="sentence-transformers/all-mpnet-base-v2")
text_embedder.warm_up()

document_embedder = OptimumDocumentEmbedder(model="sentence-transformers/all-mpnet-base-v2")
document_embedder.warm_up()

Step 4: Install and Set Up Pgvector

To quickly set up a PostgreSQL database with pgvector, you can use Docker:

docker run -d -p 5432:5432 -e POSTGRES_USER=postgres -e POSTGRES_PASSWORD=postgres -e POSTGRES_DB=postgres ankane/pgvector

To use pgvector with Haystack, install the pgvector-haystack integration:

pip install pgvector-haystack

import os
from haystack_integrations.document_stores.pgvector import PgvectorDocumentStore
from haystack_integrations.components.retrievers.pgvector import PgvectorEmbeddingRetriever

os.environ["PG_CONN_STR"] = "postgresql://postgres:postgres@localhost:5432/postgres"

document_store = PgvectorDocumentStore()
retriever = PgvectorEmbeddingRetriever(document_store=document_store)

Step 5: Build a RAG Chatbot

Now that you’ve set up all components, let’s start to build a simple chatbot. We’ll use the Milvus introduction doc as a private knowledge base. You can replace it your own dataset to customize your RAG chatbot.

url = 'https://raw.githubusercontent.com/milvus-io/milvus-docs/refs/heads/v2.5.x/site/en/about/overview.md'
example_file = 'example_file.md'
response = requests.get(url)
with open(example_file, 'wb') as f:
    f.write(response.content)
file_paths = [example_file]  # You can replace it with your own file paths.

indexing_pipeline = Pipeline()
indexing_pipeline.add_component("converter", MarkdownToDocument())
indexing_pipeline.add_component("splitter", DocumentSplitter(split_by="sentence", split_length=2))
indexing_pipeline.add_component("embedder", document_embedder)
indexing_pipeline.add_component("writer", DocumentWriter(document_store))
indexing_pipeline.connect("converter", "splitter")
indexing_pipeline.connect("splitter", "embedder")
indexing_pipeline.connect("embedder", "writer")
indexing_pipeline.run({"converter": {"sources": file_paths}})

# print("Number of documents:", document_store.count_documents())
  
question = "What is Milvus?"  # You can replace it with your own question.

retrieval_pipeline = Pipeline()
retrieval_pipeline.add_component("embedder", text_embedder)
retrieval_pipeline.add_component("retriever", retriever)
retrieval_pipeline.connect("embedder", "retriever")

retrieval_results = retrieval_pipeline.run({"embedder": {"text": question}})

# for doc in retrieval_results["retriever"]["documents"]:
#     print(doc.content)
#     print("-" * 10)

from haystack.utils import Secret
from haystack.components.builders import PromptBuilder

retriever = PgvectorEmbeddingRetriever(document_store=document_store) 
 text_embedder = OptimumTextEmbedder(model="sentence-transformers/all-mpnet-base-v2")
text_embedder.warm_up()

prompt_template = """Answer the following query based on the provided context. If the context does
                     not include an answer, reply with 'I don't know'.\n
                     Query: {{query}}
                     Documents:
                     {% for doc in documents %}
                        {{ doc.content }}
                     {% endfor %}
                     Answer: 
                  """

rag_pipeline = Pipeline()
rag_pipeline.add_component("text_embedder", text_embedder)
rag_pipeline.add_component("retriever", retriever)
rag_pipeline.add_component("prompt_builder", PromptBuilder(template=prompt_template))
rag_pipeline.add_component("generator", generator)
rag_pipeline.connect("text_embedder.embedding", "retriever.query_embedding")
rag_pipeline.connect("retriever.documents", "prompt_builder.documents")
rag_pipeline.connect("prompt_builder", "generator")

results = rag_pipeline.run({"text_embedder": {"text": question}, "prompt_builder": {"query": question},})
print('RAG answer:\n', results["generator"]["replies"][0])

Optimization Tips

As you build your RAG system, optimization is key to ensuring peak performance and efficiency. While setting up the components is an essential first step, fine-tuning each one will help you create a solution that works even better and scales seamlessly. In this section, we’ll share some practical tips for optimizing all these components, giving you the edge to build smarter, faster, and more responsive RAG applications.

Haystack optimization tips

To optimize Haystack in a RAG setup, ensure you use an efficient retriever like FAISS or Milvus for scalable and fast similarity searches. Fine-tune your document store settings, such as indexing strategies and storage backends, to balance speed and accuracy. Use batch processing for embedding generation to reduce latency and optimize API calls. Leverage Haystack's pipeline caching to avoid redundant computations, especially for frequently queried documents. Tune your reader model by selecting a lightweight yet accurate transformer-based model like DistilBERT to speed up response times. Implement query rewriting or filtering techniques to enhance retrieval quality, ensuring the most relevant documents are retrieved for generation. Finally, monitor system performance with Haystack’s built-in evaluation tools to iteratively refine your setup based on real-world query performance.

pgvector optimization tips

To optimize pgvector in a Retrieval-Augmented Generation (RAG) setup, consider indexing your vectors using GiST or IVFFlat to significantly speed up search queries and improve retrieval performance. Make sure to leverage parallelization for query execution, allowing multiple queries to be processed simultaneously, especially for large datasets. Optimize memory usage by tuning the vector storage size and using compressed embeddings where possible. To further enhance query speed, implement pre-filtering techniques to narrow down search space before querying. Regularly rebuild indexes to ensure they are up to date with any new data. Fine-tune vectorization models to reduce dimensionality without sacrificing accuracy, thus improving both storage efficiency and retrieval times. Finally, manage resource allocation carefully, utilizing horizontal scaling for larger datasets and offloading intensive operations to dedicated processing units to maintain responsiveness during high-traffic periods.

STACKIT Meta-Llama-3.1-70B-Instruct-FP8 optimization tips

To optimize STACKIT Meta-Llama-3.1-70B-Instruct-FP8 in RAG, prioritize reducing input token length by truncating or summarizing retrieved documents to fit the model’s context window. Use FP8 precision to accelerate inference while maintaining accuracy. Batch processing for parallel queries improves throughput. Cache frequent or repetitive queries to avoid redundant computations. Fine-tune retrieval thresholds to balance relevance and noise. Leverage hardware-specific optimizations (e.g., Tensor Cores on GPUs) and enable model parallelism for distributed workloads. Monitor memory usage to prevent bottlenecks, and prune low-confidence outputs early to save resources.

Optimum all-mpnet-base-v2 optimization tips

To optimize Optimum all-mpnet-base-v2 in a RAG setup, preprocess input text by trimming redundant whitespace, normalizing casing, and splitting long documents into smaller chunks (≤512 tokens) to align with the model’s max sequence length. Use batch processing for embeddings to leverage GPU parallelism, adjusting batch size based on GPU memory. Quantize the model via ONNX Runtime or FP16 precision for faster inference. Cache frequently accessed embeddings to reduce recomputation, and pair with efficient vector search libraries (e.g., FAISS) for low-latency retrieval. Regularly update and prune the document corpus to maintain relevance.

By implementing these tips across your components, you'll be able to enhance the performance and functionality of your RAG system, ensuring it’s optimized for both speed and accuracy. Keep testing, iterating, and refining your setup to stay ahead in the ever-evolving world of AI development.

RAG Cost Calculator: A Free Tool to Calculate Your Cost in Seconds

Estimating the cost of a Retrieval-Augmented Generation (RAG) pipeline involves analyzing expenses across vector storage, compute resources, and API usage. Key cost drivers include vector database queries, embedding generation, and LLM inference.

RAG Cost Calculator is a free tool that quickly estimates the cost of building a RAG pipeline, including chunking, embedding, vector storage/search, and LLM generation. It also helps you identify cost-saving opportunities and achieve up to 10x cost reduction on vector databases with the serverless option.

Calculate your RAG cost now.

Calculate your RAG cost

What Have You Learned?

By diving into this tutorial, you’ve unlocked the magic of building a powerful RAG system from the ground up! You’ve seen firsthand how Haystack acts as the backbone, seamlessly orchestrating the flow of data between components, while Pgvector transforms PostgreSQL into a lightning-fast vector database perfect for storing and retrieving embeddings. The Optimum all-mpnet-base-v2 model showed off its chops as a top-tier embedding engine, converting text into rich numerical representations that capture meaning with precision. And when it came time to generate answers, the STACKIT Meta-Llama-3.1-70B-Instruct-FP8 model stepped into the spotlight, delivering human-like responses that feel both accurate and natural. Together, these tools form a dynamic pipeline that can tackle everything from complex queries to creative tasks—proving that cutting-edge AI isn’t just for big tech companies anymore. You’ve also learned practical optimization tricks, like tuning retrieval parameters and balancing speed with accuracy, to make your system leaner and smarter. Plus, the free RAG cost calculator gave you a clear roadmap for budgeting resources without sacrificing performance.

Now that you’ve seen how these pieces fit together, imagine the possibilities! Whether you’re building a chatbot, enhancing search, or crafting a knowledge hub, you’ve got the tools to innovate. The tutorial didn’t just teach you how to connect frameworks and models—it showed you how to think like an architect, blending creativity with technical rigor. Don’t stop here! Experiment with different datasets, tweak your embeddings, or swap in new LLMs to see how your system evolves. The world of RAG is wide open, and you’re equipped to push boundaries. So fire up your IDE, spin up a database, and start building something that amazes. Your next breakthrough is just a few lines of code away—let’s make it happen! 🚀

Further Resources

🌟 In addition to this RAG tutorial, unleash your full potential with these incredible resources to level up your RAG skills.

How to Build a Multimodal RAG | Documentation
How to Enhance the Performance of Your RAG Pipeline
Graph RAG with Milvus | Documentation
How to Evaluate RAG Applications - Zilliz Learn
Generative AI Resource Hub | Zilliz

We'd Love to Hear What You Think!

We’d love to hear your thoughts! 🌟 Leave your questions or comments below or join our vibrant Milvus Discord community to share your experiences, ask questions, or connect with thousands of AI enthusiasts. Your journey matters to us!

If you like this tutorial, show your support by giving our Milvus GitHub repo a star ⭐—it means the world to us and inspires us to keep creating! 💖