DEV Community

Cover image for Try AnimeGANv2 with PyTorch on Google Colab
0xkoji
0xkoji

Posted on

Try AnimeGANv2 with PyTorch on Google Colab

What is AnimeGANV2?

AnimeGANv2, the improved version of AnimeGAN.
AnimeGAN is a lightweight GAN for a photo animation.

In brief, people can generate a photo that looks like an animation's scene from an image.

You can try AnimeGAN easily with this.
https://animegan.js.org/

Detail
https://tachibanayoshino.github.io/AnimeGANv2/

repos

Original repo

GitHub logo TachibanaYoshino / AnimeGANv2

[Open Source]. The improved version of AnimeGAN. Landscape photos/videos to anime

AnimeGANv2

The improved version of AnimeGAN.

Project Page | Landscape photos / videos to anime


News

  • (2022.08.03) Added the AnimeGANv2 Colab: 🖼️ Photos Photos Colab | 🎞️ Videos Colab for videos
  • (2021.12.25) AnimeGANv3 has been released. 🎄
  • (2021.02.21) The pytorch version of AnimeGANv2 has been released, Be grateful to @bryandlee for his contribution.
  • (2020.12.25) AnimeGANv3 will be released along with its paper in the spring of 2021.

Focus:

Anime style Film Picture Number Quality Download Style Dataset
Miyazaki Hayao The Wind Rises 1752 1080p Link
Makoto Shinkai Your Name & Weathering with you 1445 BD
Kon Satoshi Paprika 1284 BDRip

News:

The improvement directions of AnimeGANv2 mainly include the following 4 points:  
Enter fullscreen mode Exit fullscreen mode
  • 1. Solve the problem of high-frequency artifacts in the generated image.

  • 2. It is easy to train and directly achieve the effects in the paper.

  • 3. Further reduce the number of parameters of the generator network. (generator size: 8.17 Mb)

PyTorch Implementation

GitHub logo bryandlee / animegan2-pytorch

PyTorch implementation of AnimeGANv2

PyTorch Implementation of AnimeGANv2

Updates

Basic Usage

Inference

python test.py --input_dir [image_folder_path] --device [cpu/cuda]

Torch Hub Usage

You can load the model via torch.hub:

import torch
model = torch.hub.load("bryandlee/animegan2-pytorch", "generator").eval()
out = model(img_tensor)  # BCHW tensor
Enter fullscreen mode Exit fullscreen mode

Currently, the following pretrained shorthands are available:

model = torch.hub.load("bryandlee/animegan2-pytorch:main", "generator", pretrained="celeba_distill")
model = torch.hub.load("bryandlee/animegan2-pytorch:main", "generator", pretrained="face_paint_512_v1")
model = torch.hub.load("bryandlee/animegan2-pytorch:main", "generator", pretrained="face_paint_512_v2")
model = torch.hub.load
Enter fullscreen mode Exit fullscreen mode

steps

  1. Create a new note on Google Colab
  2. Upload an input image
  3. Run the following code


from PIL import Image
import torch
import IPython
from IPython.display import display

# https://github.com/bryandlee/animegan2-pytorch
# load models
model_celeba = torch.hub.load("bryandlee/animegan2-pytorch:main", "generator", pretrained="celeba_distill")
model_facev1 = torch.hub.load("bryandlee/animegan2-pytorch:main", "generator", pretrained="face_paint_512_v1")
model_facev2 = torch.hub.load("bryandlee/animegan2-pytorch:main", "generator", pretrained="face_paint_512_v2")
model_paprika = torch.hub.load("bryandlee/animegan2-pytorch:main", "generator", pretrained="paprika")

face2paint = torch.hub.load("bryandlee/animegan2-pytorch:main", "face2paint", size=512)

INPUT_IMG = "sg.jpg" # input_image jpg/png 
img = Image.open(INPUT_IMG).convert("RGB")

out_celeba = face2paint(model_celeba, img)
out_facev1 = face2paint(model_facev1, img)
out_facev2 = face2paint(model_facev2, img)
out_paprika = face2paint(model_paprika, img)

# save images
out_celeba.save("out_celeba.jpg")
out_facev1.save("out_facev1.jpg")
out_facev2.save("out_facev2.jpg")
out_paprika.save("out_paprika.jpg")

# display images
display(img)
display(out_celeba)
display(out_facev1)
display(out_facev2)
display(out_paprika)


Enter fullscreen mode Exit fullscreen mode

[original]
original

[celeba_distill]
celeba_distill

[face_paint_512_v1]
face_paint_512_v1

[face_paint_512_v2]
face_paint_512_v2

[paprika]
paprika

Top comments (0)