With the release of DeepSeek R1, there is a buzz in the AI community. The open-source model offers some best-in-class performance across many metrics, even at par with state-of-the-art proprietary models in many cases. Such huge success invites attention and curiosity to learn more about it. In this article, we will look into implementing a Retrieval-Augmented Generation (RAG) system using DeepSeek R1. We will cover everything from setting up your environment to running queries with additional explanations and code snippets.As already widespread, RAG combines the strengths of retrieval-based and generation-based approaches. It retrieves relevant information from a knowledge base and uses it to generate accurate and contextually relevant responses to user queries.Some prerequisites for running the codes in this tutorial are as follows:Python installed (preferably version 3.7 or higher).Ollama installed: This framework allows running models like DeepSeek R1 locally.Now, lets look into step-by-step implementation:Step 1: Install OllamaFirst, install Ollama by following the instructions on their website. Once installed, verify the installation by running:# bashollama --versionStep 2: Run DeepSeek R1 ModelTo start the DeepSeek R1 model, open your terminal and execute:# bashollama run deepseek-r1:1.5bThis command initializes the 1.5 billion parameter version of DeepSeek R1, which is suitable for various applications.Step 3: Prepare Your Knowledge BaseA retrieval system requires a knowledge base from which it can pull information. This can be a collection of documents, articles, or any text data relevant to your domain.3.1 Load Your DocumentsYou can load documents from various sources, such as text files, databases, or web scraping. Heres an example of loading text files:# pythonimport osdef load_documents(directory): documents = [] for filename in os.listdir(directory): if filename.endswith('.txt'): with open(os.path.join(directory, filename), 'r') as file: documents.append(file.read()) return documentsdocuments = load_documents('path/to/your/documents')Step 4: Create a Vector Store for RetrievalTo enable efficient retrieval of relevant documents, you can use a vector store like FAISS (Facebook AI Similarity Search). This involves generating embeddings for your documents.4.1 Install Required LibrariesYou may need to install additional libraries for embeddings and FAISS:# bashpip install faiss-cpu huggingface-hub4.2 Generate Embeddings and Set Up FAISSHeres how to generate embeddings and set up the FAISS vector store:# pythonfrom huggingface_hub import HuggingFaceEmbeddingsimport faissimport numpy as np# Initialize the embeddings modelembeddings_model = HuggingFaceEmbeddings()# Generate embeddings for all documentsdocument_embeddings = [embeddings_model.embed(doc) for doc in documents]document_embeddings = np.array(document_embeddings).astype('float32')# Create FAISS indexindex = faiss.IndexFlatL2(document_embeddings.shape[1]) # L2 distance metricindex.add(document_embeddings) # Add document embeddings to the indexStep 5: Set Up the RetrieverYou must create a retriever based on user queries to fetch the most relevant documents.# pythonclass SimpleRetriever: def __init__(self, index, embeddings_model): self.index = index self.embeddings_model = embeddings_model def retrieve(self, query, k=3): query_embedding = self.embeddings_model.embed(query) distances, indices = self.index.search(np.array([query_embedding]).astype('float32'), k) return [documents[i] for i in indices[0]]retriever = SimpleRetriever(index, embeddings_model)Step 6: Configure DeepSeek R1 for RAGNext, a prompt template will be set up to instruct DeepSeek R1 to respond based on retrieved context.# pythonfrom ollama import Ollamafrom string import Template# Instantiate the modelllm = Ollama(model="deepseek-r1:1.5b")# Craft the prompt template using string. Template for better readabilityprompt_template = Template("""Use ONLY the context below.If unsure, say "I don't know".Keep answers under 4 sentences.Context: $contextQuestion: $questionAnswer:""")Step 7: Implement Query Handling FunctionalityNow, you can create a function that combines retrieval and generation to answer user queries:# pythondef answer_query(question): # Retrieve relevant context from the knowledge base context = retriever.retrieve(question) # Combine retrieved contexts into a single string (if multiple) combined_context = "n".join(context) # Generate an answer using DeepSeek R1 with the combined context response = llm.generate(prompt_template.substitute(context=combined_context, question=question)) return response.strip()Step 8: Running Your RAG SystemYou can now test your RAG system by calling the `answer_query` function with any question about your knowledge base.# pythonif __name__ == "__main__": user_question = "What are the key features of DeepSeek R1?" answer = answer_query(user_question) print("Answer:", answer)Access the Colab Notebook with the Complete codeIn conclusion, following these steps, you can successfully implement a Retrieval-Augmented Generation (RAG) system using DeepSeek R1. This setup allows you to retrieve information from your documents effectively and generate accurate responses based on that information. Also, explore the potential of the DeepSeek R1 model for your specific use case through this.Sources Asif RazzaqAsif Razzaq is the CEO of Marktechpost Media Inc.. As a visionary entrepreneur and engineer, Asif is committed to harnessing the potential of Artificial Intelligence for social good. His most recent endeavor is the launch of an Artificial Intelligence Media Platform, Marktechpost, which stands out for its in-depth coverage of machine learning and deep learning news that is both technically sound and easily understandable by a wide audience. The platform boasts of over 2 million monthly views, illustrating its popularity among audiences. Meet 'Height':The only autonomous project management tool (Sponsored)