Skip to content
Faceswap with Pytorch or DeepFake with Pytorch
Branch: master
Clone or download
Latest commit 8a4aae0 Jan 1, 2019
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
.idea content added Apr 16, 2018
README.md Update README.md Jan 1, 2019
Screenshot from 2018-04-16 15-36-47.png Add files via upload Apr 16, 2018
image_augmentation.py content added Apr 16, 2018
models.py Pytorch-v0.4 is supported Jul 10, 2018
padding_same_conv.py Pytorch-v0.4 is supported Jul 10, 2018
train.py some annotation added Jul 12, 2018
training_data.py content added Apr 16, 2018
umeyama.py content added Apr 16, 2018
util.py Pytorch-v0.4 is supported Jul 10, 2018

README.md

faceswap-pytorch

Deepfakes_Faceswap using pytorch JUST FOR STUDY AND RESEARCH

This is pytorch version compared with https://github.com/joshua-wu/deepfakes_faceswap which using Keras.

Processing image

Source code you can download directly from the github page.

Source code,training images and trained model(~300MB):

https://pan.baidu.com/s/197RIMB_Po96RZNFzV-wqjA PW : z4wa

Requirement:

Python == 3.6
pytorch >= 0.4.0

pytorch-1.0 is supported. You need a modern GPU and CUDA support for better performance.

How to run:

python train.py for simple run

if you don't use trained model,only after about 1000 epoch can you see the result and after 10000 epoch the result is the same with the above picture.

ps : The Pytorch trains a little faster than Keras using tf backend.

You can’t perform that action at this time.