https://github.com/p1atdev/mangalineextraction-hf

MangaLineExtraction in transformers library format
https://github.com/p1atdev/mangalineextraction-hf

Last synced: 6 days ago
JSON representation

MangaLineExtraction in transformers library format

• Host: GitHub
• URL: https://github.com/p1atdev/mangalineextraction-hf
• Owner: p1atdev
• License: apache-2.0
• Created: 2024-02-21T04:43:09.000Z (11 months ago)
• Default Branch: main
• Last Pushed: 2024-02-21T11:35:12.000Z (11 months ago)
• Last Synced: 2024-11-22T00:47:59.888Z (about 2 months ago)
• Language: Jupyter Notebook
• Homepage: https://huggingface.co/p1atdev/MangaLineExtraction-hf
• Size: 2.06 MB
• Stars: 0
• Watchers: 2
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# MangaLineExtraction-hf

A huggingface transformers compatible implementation of MangaLineExtraction (https://github.com/ljsabc/MangaLineExtraction_PyTorch).

The converted weights are avaiable on [🤗 HuggingFace](https://huggingface.co/p1atdev/MangaLineExtraction-hf).

## Example usage with transformers

```py

from PIL import Image

import torch

from transformers import AutoModel, AutoImageProcessor

REPO_NAME = "p1atdev/MangaLineExtraction-hf"

model = AutoModel.from_pretrained(REPO_NAME, trust_remote_code=True)

processor = AutoImageProcessor.from_pretrained(REPO_NAME, trust_remote_code=True)

image = Image.open("./sample.jpg")

inputs = processor(image, return_tensors="pt")

with torch.no_grad():

    outputs = model(inputs.pixel_values)

line_image = Image.fromarray(outputs.pixel_values[0].numpy().astype("uint8"), mode="L")

line_image.save("./line_image.png")

```

## Acknowledgements

We extend our gratitude to the authors of the  [Deep Extraction of Manga Structural Lines](https://www.cse.cuhk.edu.hk/~ttwong/papers/linelearn/linelearn.html) and the contributors to the  [MangaLineExtraction_PyTorch](https://github.com/ljsabc/MangaLineExtraction_PyTorch) for their pioneering work that served as the foundation for our adaptation. Our thanks also go to HuggingFace for developing [transformers](https://github.com/huggingface/transformers), enabling us to enhance this project further.

For detailed information, please refer to:

- https://www.cse.cuhk.edu.hk/~ttwong/papers/linelearn/linelearn.html

- https://github.com/huggingface/transformers