Building Chatbot that answers medical questions using the open biomriage of the Biomtral LLM, Langchain, thermaker storage, and rag: Step Guide for action

In this lesson, we will build a strong, powerful Chatbot, based on Chatbot related Chatbot with the health-related content. We will improve the skills of the open Bioming of the LLM and Langchain skills and the data variables of processing PDF documents into chunks of portable text. We will then say that these chunks use facial robbery, kidnapping the deepest semantic relationships and keep them in the chroma vector database to get the higher return. Finally, through system restoration system (RAG), we will unite the re-restored status in our Chatbot repairs, confirming clear, approved users' answers. This method allows us to stay fast through large volumes of medical PDFs, providing rich, accurate understanding, and accurate understanding.

To Set Tools

!pip install langchain sentence-transformers chromadb llama-cpp-python langchain_community pypdf
from langchain_community.document_loaders import PyPDFDirectoryLoader
from langchain.text_splitter import CharacterTextSplitter,RecursiveCharacterTextSplitter
from langchain_community.embeddings import HuggingFaceEmbeddings
from langchain.vectorstores import FAISS, Chroma
from langchain_community.llms import LlamaCpp
from langchain.chains import RetrievalQA, LLMChain
import pathlib
import textwrap
from IPython.display import display
from IPython.display import Markdown


def to_markdown(text):
    text = text.replace('•', '  *')
    return Markdown(textwrap.indent(text, '> ', predicate=lambda _: True))
from google.colab import drive
drive.mount('/content/drive')

First, we enter and prepare for Python Packages by processing documents, generating generation, local llms, and performance-based travel. We receive Langchain_Community Loading PDF and text classification, set Retrivalqa and LLMChACHAIN ​​to receive a question response, and including the use of to_markdown use

To set access to API key

from google.colab import userdata
# Or use `os.getenv('HUGGINGFACEHUB_API_TOKEN')` to fetch an environment variable.
import os
from getpass import getpass


HF_API_KEY = userdata.get("HF_API_KEY")
os.environ["HF_API_KEY"] = "HF_API_KEY"

Here, we can safely download and place Hugging Face API key such as natural flexibility in Google Colab. There can also be a natural variability of HuggingFACTACLOBLUBLOBLO'P_TOKEN avoiding direct disclosure of sensitive credentials in your code.

Loading and issuing PDFs from indicator

loader = PyPDFDirectoryLoader('/content/drive/My Drive/Data')
docs = loader.load()

We use the PypdfddfddfdictoryLler.

Divideing text documents included into unoccupied chunks

text_splitter = RecursiveCharacterTextSplitter(chunk_size=300, chunk_overlap=50)
chunks = text_splitter.split_documents(docs)

At this Code Snippet, RecsususususususususususCracRacreXplotter is used to break down each document in the documents small, more controlled parts.

Getting started binding the face of the face

embeddings = HuggingFaceEmbeddings(model_name="BAAI/bge-base-en-v1.5")

Using HuggingBanteds, we create an item that uses the BAA / BGE-BA-V1.5 model. Converts text into numerical fields.

To create a vector store and use the same search

vectorstore = Chroma.from_documents(chunks, embeddings)
query = "who is at risk of heart disease"
search = vectorstore.similarity_search(query)
to_markdown(search[0].page_content)

We begin to create a Chroma Vector store (chroma.fromom_dwoy) from Chenks Trunks and a defined model. Next, creates a question to ask, “Who is in danger of heart disease,” and do the same search against the reserves. The main result (searches[0].page_incontent) and modify in Markdown for clear display.

Creating Recovery and Download Due Documents

retriever = vectorstore.as_retriever(
    search_kwargs={'k': 5}
)
retriever.get_relevant_documents(query)

We are converting the Chroma Vector shop into retriev (vectorstore.astore.as_retetriever) well-interested in the correct documentation of the question provided.

The Department Biomstral-7b Model with LLAMCPP

llm = LlamaCpp(
    model_path= "/content/drive/MyDrive/Model/BioMistral-7B.Q4_K_M.gguf",
    temperature=0.3,
    max_tokens=2048,
    top_p=1)

We set up a Biomouts of the open local Biomistism using the Llamacpp, points to the previously downloaded model file. We also prepare for the springs of the heat, max_tokens, and top_p, control, high tokens produced, and the nucleus lampling.

Setting Up Development for Restoring Power (RAG) Chain with Quick Commorem

from langchain.schema.runnable import RunnablePassthrough
from langchain.schema.output_parser import StrOutputParser
from langchain.prompts import ChatPromptTemplate
template = """
<|context|>
You are an AI assistant that follows instruction extremely well.
Please be truthful and give direct answers

<|user|>
{query}

 <|assistant|>
"""
prompt = ChatPromptTemplate.from_template(template)
rag_chain = (
    {'context': retriever, 'query': RunnablePassthrough()}
    | prompt
    | llm
    | StrOutputParser()
)

Using the above, we put the rag pipe using the langchain frame. It creates customization by instructions and local edits, including retriever for context, and the language model to produce the answers. The flow is defined as a series of direct query (illegal chatpromortemplevate for speedy construction, llm for response to generation, and ultimately, the stroutoutputparter to produce a clean text.

Asking Rag Chain to answer the health-related question

response = rag_chain.invoke("Why should I care about my heart health?")
to_markdown(response)

Now, we call a previously designed rag. Passes the question to returns, returns the correct context from the collection of documents, and feeds the context in the LLM to produce a short, accurate indicator.

In conclusion, by combining a biomcpt with LMAMCPP and applying the Langchain variability, we are able to build medical chatbot for context. From a chunk based index based on rag-free rag pipes, directing the highest volume process of the PDF Data for the right understanding. Users find clear and read answers in format of the last answers for Markdown. The project can be expanded or made for different backgrounds, to ensure the stability and accuracy of information across different documents.

Use Brochure in the corner here. Also, don't forget to follow Sane and join ours Telegraph station including LinkedIn Grtopic. Don't forget to join ours 75k + ml subreddit.

🚨 Meet the Work: an open source opened with multiple sources to check the difficult program AI ^(Updated)