Author: Aniket Thomas (AniTho)
This consists of an API built using fastapi to use AnimeGan
Run the main file and call the api from your application through the path (<ip-address of api>/api/trans/) and pass the following inputs.
Input:
- Image Byte: Passed as a file upload
Passes the Image on which AnimeGan Model has to be used
- model_name: Passed as a json request
Specifies the weights to be used for inference.
Supports 'celeba', 'facepaintv1', 'facepaitv2', 'hayao', 'paprika', 'shinkai'
The api returns a byte object of the transformed image as a response.
Returns:
Returns a byte object of the transformed image after applying the animegan model
Note: A sample Use case is shown in test.ipynb notebook.