Skip to content

🤗🖼️ HuggingPics: Fine-tune Vision Transformers for anything using images found on the web.

Notifications You must be signed in to change notification settings

nateraw/huggingpics

Repository files navigation

🤗🖼️HuggingPics

Fine-tune Vision Transformers for anything using images found on the web.

Check out the video below for a walkthrough of this project! ⤵️

IMAGE ALT TEXT

Usage

Click on the link below to try it out:

Open In Colab

How does it work?

1. You define your search terms

pick search terms

2. We download ~150 images for each and use them to fine-tune a ViT

image search results

3. You push your model to HuggingFace's Hub to share your results with the world

push to hub

Your auto-generated model repo will look something like this. Pretty cool, eh? 😎

push to hub

Examples

💡 If you need some inspiration, take a look at the examples below:

nateraw/rare-puppers nateraw/pasta-pizza-ravioli nateraw/baseball-stadium-foods nateraw/denver-nyc-paris
term_1 samoyed pizza cotton candy denver
term_2 shiba inu pasta hamburger new york city
term_3 corgi ravioli hot dog paris
term_4 nachos
term_5 popcorn

You can see a full list of model repos created using this tool by clicking here