50f892d57d
all models: removed options 'src_scale_mod', and 'sort samples by yaw as target' If you want, you can manually remove unnecessary angles from src faceset after sort by yaw.
Colombo2019-12-21 23:16:55 +04:00
8866dce22e
added 4.2.other) data_src util faceset metadata save.bat saves metadata of data_src\aligned\ faces into data_src\aligned\meta.dat
Colombo2019-12-20 15:03:17 +04:00
cf27558e88
SAEHD: lr_dropout now as an option, and disabled by default
Colombo2019-12-20 10:39:05 +04:00
64021b9c62
more stable and precise version of face transformation matrix. fixed bleeding mask on some samples
Colombo2019-12-20 10:30:49 +04:00
068c7d0d55
temporary revert last fixes
Colombo2019-12-20 10:21:59 +04:00
dd1d5e8909
improved face align, More stable and precise version of the face transformation matrix. Now full_faces are aligned with the upper and lateral boundaries of the frame, result: fix of cutted mouth, increase area of the cheeks of side faces before/after https://i.imgur.com/t9IyGZv.jpg therefore, additional training is required for existing models. Optionally, you can re-extract dst faces of your project, if they have problems with cutted mouth or cheeks.
Colombo2019-12-19 18:33:04 +04:00
28b948888c
fix extractor for some systems
Colombo2019-12-17 09:52:48 +04:00
71ebf06c89
SAEHD,Quick96: improved model generalization, overall accuracy and sharpness by using new 'Learning rate dropout' technique from paper https://arxiv.org/abs/1912.00144 An example of a loss histogram where this function is enabled after the red arrow: https://i.imgur.com/3olskOd.jpgColombo2019-12-15 15:53:06 +04:00
d4745b5cf8
added sort by absdiff This is sort method by absolute per pixel difference between all faces. options: Sort by similar? ( y/n ?:help skip:y ) : if you choose 'n', then most dissimilar faces will be placed first.
Colombo2019-12-11 22:33:49 +04:00
c0f258c336
SAE,SAEHD,Converter: added sot-m color transfer
Colombo2019-11-12 09:07:50 +04:00
05153d9ba5
FacesetRelighter fixes and improvements:
Colombo2019-11-11 19:56:36 +04:00
fe58459f36
added FacesetRelighter: Synthesize new faces from existing ones by relighting them using DeepPortraitRelighter network. With the relighted faces neural network will better reproduce face shadows.
Colombo2019-11-11 11:42:52 +04:00
b9c0815d17
added random shuffle to samples in SampleGeneratorFace.py
Colombo2019-11-11 11:35:46 +04:00
f6f8f8ee5d
change default jpg face extractor quality to 100
Colombo2019-11-11 09:13:42 +04:00
7d389718fe
added Quick96 model This is the fastest model for low-end cards. Model has zero options and trains a 96pix fullface. It is good for quick deepfake demo. Example of the preview trained in 15 minutes on RTX2080Ti: https://i.imgur.com/oRMvZFP.jpgColombo2019-11-09 19:24:31 +04:00
59d6fada23
draw up arrow in the red landmark debug square
Colombo2019-10-24 10:09:51 +04:00
734d97d729
added 'sort by vggface': sorting by face similarity using VGGFace model. Requires 4GB+ VRAM and internet connection for the first run.
Colombo2019-10-23 15:06:39 +04:00
d3e35c8ce6
enhanced sort by final: now faces are evenly distributed not only in the direction of yaw, but also in pitch
Colombo2019-10-20 14:18:14 +04:00
92f14dee70
SAEHD: added option Enable random warp of samples, default is on Random warp is required to generalize facial expressions of both faces. When the face is trained enough, you can disable it to get extra sharpness for less amount of iterations.
Colombo2019-10-12 10:31:50 +04:00
e15f846d08
SAE, SAEHD: random flip and learn mask options now can be overridden
Colombo2019-10-10 22:56:57 +04:00
1f27d13f61
clip border for midface
Colombo2019-10-10 14:48:51 +04:00
63ecd4ec34
speed up S3FD extraction for GPU with >= 11GB vram
Colombo2019-10-10 13:22:14 +04:00
d781af3d1f
fixed GPU detection and indexes, got rid of using nvml, now using direct cuda lib to determine gpu info that match tensorflow indexes,
Colombo2019-10-05 16:26:23 +04:00
Colombo
