https://github.com/daspartho/is-it-huggable

An image classifier to classify things as huggable or not.
https://github.com/daspartho/is-it-huggable

ai deep-learning fastai huggingface-spaces image-classification jupyter-notebook python

Last synced: 6 days ago
JSON representation

An image classifier to classify things as huggable or not.

• Host: GitHub
• URL: https://github.com/daspartho/is-it-huggable
• Owner: daspartho
• License: apache-2.0
• Created: 2022-08-23T17:23:03.000Z (over 2 years ago)
• Default Branch: main
• Last Pushed: 2023-11-28T10:35:01.000Z (about 1 year ago)
• Last Synced: 2024-09-27T01:54:23.800Z (3 months ago)
• Topics: ai, deep-learning, fastai, huggingface-spaces, image-classification, jupyter-notebook, python
• Language: Jupyter Notebook
• Homepage: https://daspartho.github.io/is-it-huggable/
• Size: 1.77 MB
• Stars: 28
• Watchers: 1
• Forks: 0
• Open Issues: 2
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

• awesome-useful-projects - <img src="https://icon.horse/icon/github.com" height="20px" align="center"/>/daspartho/is-it-huggable - Analyzes images to determine if it is huggable (Jump To / Other)
• awesome-useful-projects - <img src="https://icon.horse/icon/github.com" height="20px" align="center"/>/daspartho/is-it-huggable - Analyzes images to determine if it is huggable (Jump To / Other)

README

        
# Is it Huggable?

An image classifier that tells if something's huggable or not? 

Made using [fastai](https://docs.fast.ai/).

Play with it on the [website](https://daspartho.github.io/is-it-huggable/) or on [HuggingFace Space](https://huggingface.co/spaces/daspartho/is-it-huggable).

# How it works?

A pretrained image classification model (here resnet34) is fine tuned on images of multiple examples of both categories, like for huggable, pillow photos and for not huggable, images of cactus.

For steps to make the model check the [model](https://github.com/daspartho/is-it-huggable/blob/main/model.ipynb) notebook in the repo.

# Examples

![](demo-images/plushie.png)

![](demo-images/chainsaw.png)

![](demo-images/shark.png)

![](demo-images/snowman.png)

![](demo-images/dog.png)

![](demo-images/tiger.png)

![](demo-images/cactus.png)

![](demo-images/bed.png)

![](demo-images/trex.png)

![](demo-images/knife.png)

![](demo-images/bunny.png)

# Important Notes

- Huggability is very subjective.

- The model is not looking for particular faces or characters but general features of it.

- It's just picking up on features that makes a thing huggable or vice-versa from the training data, like sharpness.

# Contributing

If you want to contribute code, just write a quick pull request and the developers will take a look at it. If you want to suggest an idea, just write an issue and the developers will check it out!