46ff33bf89
SAE: dssim kernel size now depends on resolution
iperov
2019-03-12 09:49:40 +04:00
fd3b9add2f
SAE: added option "simple optimizer" allows to train bigger networks on same VRAM nnlib: added DFLOptimizer is my own optimizer
iperov
2019-03-12 09:32:35 +04:00
ee8dbcbc35
revert back Adam
iperov
2019-03-11 21:52:36 +04:00
e4637336ef
added ability to save optimizers states which work with K.function, added custom Adam that can save 'iterations' param
iperov
2019-03-11 18:23:01 +04:00
168d48b119
extractor: increased speed of 1st pass by DLIB for >= 4GB vram cards, added 'h' key for hide the help in manual extractor
iperov
2019-03-08 17:01:11 +04:00
31c2298b5f
Converter: added option for seamless to supress jittering, Lenx,leny region now averaged by grayscale gradients, now uses all CPU. SAE: multiscale_decoder option default = False update readme and manual_ru.pdf
iperov
2019-03-03 15:33:52 +04:00
46f4d6c426
fix cpu fallback for extractor
iperov
2019-02-28 14:54:59 +04:00
8f2e69ab27
upgrading to cuda 10.0, cudnn 7.4.1, tf 1.13.1, dlib 19.16.0
iperov
2019-02-28 13:39:54 +04:00
438213e97c
manual extractor: increased FPS, sort by final : now you can specify target number of images, converter: fix seamless mask and exception, huge refactoring
iperov
2019-02-28 11:56:31 +04:00
7db469a1da
fix nnlib/device.py for some notebooks
iperov
2019-02-25 21:34:47 +04:00
daebe10ca7
Extractor: fix aspect ratio for square images, now we can re-extract face manually from already aligned jpg image without quality loss.
iperov
2019-02-22 17:28:08 +04:00
72ba6b103c
added support of AMD videocards added Intel's plaidML backend to use OpenCL engine. Check new requirements. smart choosing of backend in device.py env var 'force_plaidML' can be choosed to forced using plaidML all tf functions transferred to pure keras MTCNN transferred to pure keras, but it works slow on plaidML (forced to CPU in this case) default batch size for all models and VRAMs now 4, feel free to adjust it on your own SAE: default style options now ZERO, because there are no best values for all scenes, set them on your own. SAE: return back option pixel_loss, feel free to enable it on your own. SAE: added option multiscale_decoder default is true, but you can disable it to get 100% same as H,DF,LIAEF model behaviour. fix converter output to .png added linux fork reference to doc/doc_build_and_repository_info.md
iperov
2019-02-19 17:33:12 +04:00
3a9d450281
little fix
iperov
2019-02-16 21:14:15 +04:00
06fe1314d8
removing default yaw_value from DFLIMG files, added better pitch/yaw estimator from 68 landmarks, improving face yaw accuracy for sorting and trainers, added sort by face-pitch
iperov
2019-02-12 21:31:37 +04:00
854ab11de3
fix bug with samples that were not clipped after tanh-untanh transformations, upd README.md
iperov
2019-02-10 10:45:51 +04:00
51a13c90d1
SAE: you have to restart training, added multiscale decoder as option. mask now training as not multiscaled
iperov
2019-02-09 20:33:26 +04:00
4d37fd62cd
fix DFLJPG, SAE: added "rare sample booster" SAE: pixel loss replaced to smooth transition from DSSIM to PixelLoss in 15k epochs by default
iperov
2019-02-09 18:53:37 +04:00
iperov