RAG app with GROQ with unlimited rate limit!

Today, we’re diving into the exciting world of LangChain RAG applications, a revolutionary approach to building intelligent and informative AI experiences.

What is LangChain RAG?

Imagine an LLM that can not only access its vast internal knowledge but can also tap into the specific data you provide. That’s the magic of Retrieval-Augmented Generation (RAG) with LangChain. Here’s the gist:

• Data Retrieval: LangChain’s RAG framework lets you create a custom index of your data, be it documents, code, or any other relevant information.
• Smart Assistant: When a user interacts with your application, LangChain retrieves the most relevant data points from your index based on their query.
• Supercharged LLM: This retrieved data is then fed to the LLM, essentially giving it context and boosting its ability to understand and respond to the user’s specific needs.

Why Use LangChain RAG?

The benefits of LangChain RAG are numerous. Here are a few to get you excited:

• Domain Expertise: Inject your LLM with the knowledge of your specific field, allowing it to answer complex questions and generate human-quality text tailored to your domain.
• Private Data Power: Use your own private data that LLMs wouldn’t normally have access to, unlocking a whole new level of personalization and accuracy.
• Always Up-to-Date: Keep your LLM sharp by incorporating the latest information from your data source, ensuring your application stays relevant.

*Building Your First LangChain RAG App with no rate limit!
*
The LangChain framework offers a user-friendly platform to build your RAG application. Here’s a quick peek at the process:

• Data Preparation: Structure your data for efficient retrieval. This might involve cleaning, transforming, and indexing your information.
• LangChain Pipeline: Craft a LangChain pipeline that retrieves relevant data points based on user queries and feeds them to your chosen LLM.
• Interactive Interface: Design a user-friendly interface where users can interact with your AI-powered application.

So lets Dive into code:

You can start selecting the build in IDE, Google colab which has free RAM and GPUs in it to give you freedom for creativity in AI/ML etc.

Click the following link:

Google colab

Lets start coding!

Side by side do it with me (Press windows + left arrow key and blog on window + right arrow key)

Install the framworks from langchain in colab

! pip install langchain_community tiktoken langchain-groq langchainhub chromadb langchain langchain_core sentence-transformers pypdf

Sign up at Groq.com for your free API key!

os.environ['GROQ_API_KEY'] = '<your API key>'

Run the following code for uploading a document in colab:)

from google.colab import files
uploaded = files.upload()

RAG code for Talking to your pdf document:

import bs4
from langchain import hub
from langchain.text_splitter import RecursiveCharacterTextSplitter
from langchain_community.document_loaders import WebBaseLoader
from langchain_community.vectorstores import Chroma
from langchain_core.output_parsers import StrOutputParser
from langchain_core.runnables import RunnablePassthrough
from langchain_groq import ChatGroq
from langchain.embeddings import HuggingFaceEmbeddings
from langchain_core.embeddings import Embeddings
from langchain_community.document_loaders import TextLoader
from langchain_community.document_loaders import PyPDFLoader


#### INDEXING ####


# Load pdf
loader = PyPDFLoader("./press ctrl+space")
pages = loader.load_and_split()


# Split

text_splitter = RecursiveCharacterTextSplitter(chunk_size=1000, chunk_overlap=200)
splits = text_splitter.split_documents(pages)


# Embed

vectorstore = Chroma.from_documents(documents=splits,
                                    embedding=HuggingFaceEmbeddings())

retriever = vectorstore.as_retriever()


#### RETRIEVAL and GENERATION ####


# Prompt
prompt = hub.pull("rlm/rag-prompt")


# LLM

llm = ChatGroq(model_name="mixtral-8x7b-32768", temperature=0)


# Post-processing

def format_docs(docs):
    return "\n\n".join(doc.page_content for doc in docs)


# Chain

rag_chain = (
    {"context": retriever | format_docs, "question": RunnablePassthrough()}
    | prompt
    | llm
    | StrOutputParser()
)


# Question

rag_chain.invoke("")# write a question related to pdf data!

ooha! You completed the project of Talk to your pdfs with Retrieval Augment Generation model.

Writing blogs on open source projects!

stay tuned for the next blog, follow!

I will build next with Llama 3 and other open source model.

If you any problem you can join my discord community we build projects daily, So I named it Projects Every.

Subscribe:)