|
--- |
|
title: Action Detection In Images |
|
emoji: 😻 |
|
colorFrom: blue |
|
colorTo: pink |
|
sdk: gradio |
|
sdk_version: 4.31.5 |
|
app_file: app.py |
|
pinned: false |
|
license: mit |
|
--- |
|
|
|
Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference |
|
|
|
# General Action Classifier |
|
|
|
This is a Gradio interface that allows users to upload an image and specify candidate labels to check if a certain action is present in the image. The app uses a CLIP-ViT model to classify the image based on the provided labels. |
|
|
|
## How to Use |
|
|
|
1. Upload an image. |
|
2. Enter candidate labels separated by commas. |
|
3. The app will classify the image and display the results. |
|
|
|
## Example |
|
|
|
For instance, to check if a person is holding a beverage, you can enter labels like: "human with beverage, human, beverage". |
|
|
