Commit graph

75 commits

Author SHA1 Message Date
iperov
893354b870 fix S3FD for plaidML, fix S3FD warnings on Lambda layers. 2019-03-11 09:32:00 +04:00
iperov
fbf39d2727 added new extractor: S3FD,
all extractors now produce less false-positive faces
2019-03-10 23:18:10 +04:00
iperov
168d48b119 extractor: increased speed of 1st pass by DLIB for >= 4GB vram cards,
added 'h' key for hide the help in manual extractor
2019-03-08 17:01:11 +04:00
iperov
6094e5ead8 added Util.recover_original_aligned_filename 2019-03-04 10:57:44 +04:00
iperov
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
2019-03-03 15:33:52 +04:00
iperov
87236921a5 fix VideoEd.py 2019-03-02 18:06:06 +04:00
iperov
1d52e665f8 fix VideoEd.py 2019-03-02 18:02:00 +04:00
iperov
4196e962a8 added VideoEd - operating video files by ffmpeg from python. 2019-03-02 14:58:22 +04:00
iperov
46f4d6c426 fix cpu fallback for extractor 2019-02-28 14:54:59 +04:00
iperov
438213e97c manual extractor: increased FPS,
sort by final : now you can specify target number of images,
converter: fix seamless mask and exception,
huge refactoring
2019-02-28 11:56:31 +04:00
iperov
daebe10ca7 Extractor: fix aspect ratio for square images,
now we can re-extract face manually from already aligned jpg image without quality loss.
2019-02-22 17:28:08 +04:00
iperov
5bbd8997d9 fix Sorter.py 2019-02-22 15:22:42 +04:00
iperov
5e474c3462 added sort by one face in images. it trashes all images in which more than 1 face is found 2019-02-22 11:32:27 +04:00
iperov
f0a20b46d3 SAE: added new archi 'vg' 2019-02-21 17:53:59 +04:00
andy-ger
cd4e5b9ba6 Fixed typo (#156) 2019-02-20 18:12:30 +04:00
iperov
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
2019-02-19 17:33:12 +04:00
iperov
72c7db4980 trainer preview: added [l]: change history range 2019-02-15 19:27:44 +04:00
iperov
5a727f4e50 fix manual extractor bugs 2019-02-13 11:02:12 +04:00
iperov
6e12594af1 added util --add-landmarks-debug-images 2019-02-13 10:17:08 +04:00
iperov
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
2019-02-12 21:31:37 +04:00
iperov
e5995630fc fix 2019-02-10 23:00:08 +04:00
iperov
0866b2c509 manual extractor: added zoom for small faces 2019-02-10 22:40:57 +04:00
iperov
bda3fb3da1 refactoring 2019-02-10 21:40:58 +04:00
iperov
a65b557b0e loss history with epoch now saves to preview history images 2019-02-07 22:34:51 +04:00
iperov
0d76bee2ee fix Trainer.py for debug 2019-02-07 11:37:55 +04:00
iperov
72a640c66c fox Converter.py for debug 2019-02-07 11:29:20 +04:00
iperov
5a1fb199f4 added support of non-english characters in paths 2019-02-05 19:33:55 +04:00
iperov
883da4b9ed fix Util.py, SAE: lowest ae dims now can be 32 (for testing) 2019-02-05 14:43:54 +04:00
iperov
0fb3dd53ce fix extractor jpg quality 2019-02-05 14:08:56 +04:00
iperov
6d95dd4a99 DFL now works with JPG files by default. Old PNG files also supported. Added util convertor PNG to JPG. 2019-02-04 19:47:24 +04:00
iperov
7ed7f2bab9 fixed bug in manual extractor on windows 10 2019-01-26 06:20:47 +04:00
iperov
72646becd1 forcing all tqdm's to ascii in order to work properly on Chinese windows 2019-01-24 21:32:07 +04:00
iperov
318e6608f7 Extractor: increased speed of deleted debug files searcher 2019-01-23 09:48:45 +04:00
iperov
21460c2e13 upd manual extractor help message 2019-01-22 18:15:55 +04:00
iperov
2571f23bc4 Extractor added new feature --manual-output-debug-fix,
Performs manual reextract input-dir frames which were deleted from [output_dir]_debug\ dir.
2019-01-22 12:46:54 +04:00
iperov
b6c4171ea1 optimizations of nnlib and SampleGeneratorFace,
refactorings
2019-01-22 11:52:04 +04:00
iperov
3bbc008207 change jpeg quality to 50% 2019-01-22 10:05:20 +04:00
iperov
8016ee32cc change aligned_debug files from png to jpg 2019-01-22 09:55:41 +04:00
iperov
946688567d enhanced SAE model. You should to restart training.
new default style power = 2.0
fix DF default batch sizes.
upd readme
2019-01-17 21:41:40 +04:00
iperov
1f2b1481ef now you can train models on multiple GPU's on same workspace without cloning any folders.
Model files names will be prefixed with GPU index if GPU choosed explicitly on train/convert start.
if you leave GPU idx choice default, then best GPU idx will be choosed and model file names will not contain index prefix.
It gives you possibility to train same fake with various models or options on multiple GPUs.

H64 and H128: now you can choose 'Lighter autoencoder'. It is same as vram gb <= 4 before this update.

added archived_models.zip contains old experiments

RecycleGAN: archived

devicelib: if your system has no NVML installed (some old cards), then it will work with gpu_idx=0 as 'Generic GeForce GPU' with 2GB vram.

refactorings
2019-01-14 10:48:23 +04:00
iperov
12383570e8 Fix for systems without NVSMI 2019-01-13 16:38:17 +04:00
iperov
4625bcec1c fix multigpu extractor 2019-01-13 13:59:22 +04:00
iperov
2473dabf37 fix loss history 2019-01-11 19:19:29 +04:00
iperov
c3f175862a fix manual extractor bug 2019-01-10 05:12:02 +04:00
iperov
e8620919a7 update devicelib/nnlib to detect compute capability 2019-01-08 10:48:06 +04:00
iperov
cf41e1b962 fix sort by final 2019-01-07 17:27:10 +04:00
iperov
ff3a208e80 fix sort by final 2019-01-07 14:57:46 +04:00
iperov
abdbe7fd8d added sort by 'final' 2019-01-07 14:20:44 +04:00
iperov
b5ba7d52cb increased speed of sort by hist sim for ten thousands of faces 2019-01-04 22:54:09 +04:00
iperov
57e1e59c03 revert sort_by_hist 2019-01-02 23:53:27 +04:00