Geek Out Time: Building Your Cheap Custom Chatbot using Gradio with Fine-Tuned Models on Hugging…

Geek Out Time: Building Your Cheap Custom Chatbot using Gradio with Fine-Tuned Models on Hugging Face Spaces -Part 1 — Deploying DeepSeek-7B Chat on Hugging Face Spaces’ ZeroGPU Slot

A few days ago I set out to run the DeepSeek‑7B chat model entirely on Hugging Face Spaces’ ZeroGPU slot. ZeroGPU gives you on‑demand access to an A100 GPU for free, provided you decorate your inference function and install the right bitsandbytes wheel. In this first installment of our series on building a custom, cost‑efficient chatbot with fine‑tuned models on Hugging Face Spaces, I’ll cover:

Why ZeroGPU on Hugging Face Spaces

ZeroGPU is a managed service that allocates an A100 GPU to your Space only when you explicitly mark a function with @spaces.GPU . This means you can:

This is an excerpt — the full article continues on Medium.