Phi-3+Streamlit+Ollama : STOP PAYING for ChatGPT with this NEW & LOCAL RAG Chatbot! (PDFs, DOCs)

In this video, I'll be telling you about how you can self-host your Phi-3 mini based RAG Chatbot .It is a fully free option for those who don't want to pay for ChatGPT and are looking for a free and better alternativ to it. You can also use it with Mistral AI's Mixtral 8x22b, LLaMa, Gemini and OpenAI API.

[Key Takeaways]
• 🚀 Build a chatbot that can summarize documents and URLs using Phi-3 mini model, a free and efficient AI alternative to ChatGPT.
• 💻 Run AI models on your phone or low-end computer with Phi-3, no need for high-end GPUs or expensive subscription services.
• 📊 Use Streamlit interface combined with Phi-3, served via Ollama, to create a user-friendly chatbot experience.
• 📂 Clone the Phidata repo and follow the Streamlit cookbook to set up Phi-3 with Streamlit.
• 💸 Save money by using Phi-3, which is free and doesn't require a $20/month subscription like ChatGPT.
• 📄 Summarize documents and URLs quickly and efficiently using Phi-3, even on low-end hardware.
• 🤖 Create a virtual environment and install necessary libraries using Python and Docker to run the Streamlit app.
• 👍 Phi-3 is a great alternative to ChatGPT, offering similar functionality at no cost and with lower system requirements.

[codelucky]

I'm abandoning all local LLM due to insufficient resources. Why not use GROQ's API instead?
What about GROQ + STREAMLIT + OLLAMA (Embed) ?

[kajubarfi]

Hey nice tutorial. One Question: Can we deploy it directly on streamlit website instead of local laptop? If yes kindly guide the procedure. Thank you!

[b.]

Its cool but open WebUI is more sexy 👌

[kevinfox]

I use Open web ui but is phi3 more accurate compared to llama3 in RAG?

[wasimdorboz]

each time run it i need to repeat all steps ?