How to use Hugging Face🤗& Gradio to create simple Machine Learning (ML) projects

Access to educational resources on AI/ML is quite plentiful over the web nowadays. One can find free tutorials, courses or paid ones quite easily. While most want to dive deep into AI and understand its basic blocks, some just want to tinker around with some applications. And those who only want to just try out AI apps or demos can do so on Hugging Face🤗.

Hugging Face🤗 is amazing. It has AI models, datasets and Spaces that can host the community AI apps to try out. I personally had been trying out Stable Diffusion on Spaces, and it has been fun trying out different sets of words that identifies with me.

What is further incredible is the access to model weights! I can import models with their weights from Hugging Face. Hugging Face is like GitHub but supercharged for ML.

Going forward, I recently spoke on how we can use Hugging Face and Gradio to develop simple ML apps. The slides are available here and the demo for Object Detection with DETR can be found here. The slides contain various screen recordings on how to access Spaces on Hugging Face and the different tasks that can be accomplished with the AI models.

Here are some apps I showcased during the presentation:
A simple object detection app
A translator from MetaAI for ~200 languages
Stable Diffusion to generate images from prompts
Craiyon, another text to image generator

It stills feel unreal how you can create an ML demo in less than 50 lines of code. Just import the model from Hugging Face🤗and link it with Gradio for the GUI, and done!

The slides are available here and the demo for Object Detection with DETR can be found here.

Leave a Reply

Your email address will not be published.

%d bloggers like this: