DeepFaceLab is the leading software for creating deepfakes.
Find a file
Colombo ac7725163d removed SAEv2,
added SAEHD model ( High Definition Styled AutoEncoder )
This is a new heavyweight model for high-end cards to achieve maximum possible deepfake quality in 2020.
Differences from SAE:
+ new encoder produces more stable face and less scale jitter
  before: https://i.imgur.com/4jUcol8.gifv
  after:  https://i.imgur.com/lyiax49.gifv - scale of the face is less changed within frame size
+ new decoder produces subpixel clear result
+ pixel loss and dssim loss are merged together to achieve both training speed and pixel trueness
+ by default networks will be initialized with CA weights, but only after first successful iteration
  therefore you can test network size and batch size before weights initialization process
+ new neural network optimizer consumes less VRAM than before
+ added option <Enable 'true face' training>
  The result face will be more like src and will get extra sharpness.
  example: https://i.imgur.com/ME3A7dI.gifv
  Enable it for last 15-30k iterations before conversion.
+ encoder and decoder dims are merged to one parameter encoder/decoder dims
+ added mid-full face, which covers 30% more area than half face.
2019-10-08 15:09:28 +04:00
.github upd ISSUE_TEMPLATE.md 2019-01-24 15:15:56 +04:00
converters fixed GPU detection and indexes, got rid of using nvml, now using direct cuda lib to determine gpu info that match tensorflow indexes, 2019-10-05 16:26:23 +04:00
doc removed SAEv2, 2019-10-08 15:09:28 +04:00
ebsynth Converter: 2019-09-07 13:57:42 +04:00
facelib nothing interesting 2019-10-06 17:55:32 +04:00
imagelib Converter: added new color transfer modes: mkl, mkl-m, idt, idt-m 2019-09-07 22:01:01 +04:00
interact SAE : WARNING, RETRAIN IS REQUIRED ! 2019-09-19 11:13:56 +04:00
joblib Added interactive converter. 2019-08-24 12:57:29 +04:00
localization DockerFile for Mac users to run DeepfaceLab with CPU Mode (#95) 2019-01-01 18:08:21 +04:00
mainscripts S3FD and 2DFAN-4 were improperly ported from pytorch. now fixed. 2019-09-20 17:16:37 +04:00
mathlib removing trailing spaces 2019-03-19 23:53:27 +04:00
models removed SAEv2, 2019-10-08 15:09:28 +04:00
nnlib fixed GPU detection and indexes, got rid of using nvml, now using direct cuda lib to determine gpu info that match tensorflow indexes, 2019-10-05 16:26:23 +04:00
samplelib fixed GPU detection and indexes, got rid of using nvml, now using direct cuda lib to determine gpu info that match tensorflow indexes, 2019-10-05 16:26:23 +04:00
utils Converter: 2019-09-07 13:57:42 +04:00
.gitignore added new extractor: S3FD, 2019-03-10 23:18:10 +04:00
CODEGUIDELINES initial 2018-06-04 17:12:43 +04:00
LICENSE Initial commit 2018-06-04 17:10:00 +04:00
main.py SAE : WARNING, RETRAIN IS REQUIRED ! 2019-09-19 11:13:56 +04:00
README.md 1 2019-09-19 19:36:46 +04:00
requirements-colab.txt Added interactive converter. 2019-08-24 12:57:29 +04:00
requirements-cpu.txt Added interactive converter. 2019-08-24 12:57:29 +04:00
requirements-cuda.txt Added interactive converter. 2019-08-24 12:57:29 +04:00
requirements-opencl.txt Added interactive converter. 2019-08-24 12:57:29 +04:00

#deepfakes #faceswap #face-swap #deep-learning #deeplearning #deep-neural-networks #deepface #deep-face-swap #fakeapp #fake-app #neural-networks #neural-nets

DeepFaceLab is a tool that utilizes machine learning to replace faces in videos.

If you like this software, please consider a donation.

GOAL: next DeepFacelab update.

Donate via Yandex.Money

Donate via Paypal

bitcoin:31mPd6DxPCzbpCMZk4k1koWAbErSyqkAXr

English (google translated)

На русском

Google Colab fork by @chervonij

Linux fork by @lbfs - may be outdated

(Chinese) QQ group 951138799 for ML/AI experts

deepfakes (Chinese)

reddit r/GifFakes/ (English)

reddit r/SFWdeepfakes/ (English)