5ac7e5d7f1
changed help message for pixel loss: Pixel loss may help to enhance fine details and stabilize face color. Use it only if quality does not improve over time.
iperov2019-04-04 10:22:53 +04:00
01e98cde8e
upd pixel_loss help message
iperov2019-03-31 19:21:04 +04:00
a885838363
make seamless great again! fixed seamless face jitter removed options Suppress seamless jitter, seamless erode mask modifier. seamlessed face now properly uses blur modifier added option 'FAN-prd&dst' - using multiplied FAN prd and dst mask,
iperov2019-03-30 11:58:40 +04:00
85c01e3b4a
Converter: added Apply super resolution? (y/n skip:n) : , Enhance details by applying DCSCN network. refactorings
iperov2019-03-28 21:50:27 +04:00
4683c362ac
changing SubpixelUpscaler to variable H,W dims, tensorflow backend : using depth_to_space in SubpixelUpscaler, so training speed increased by 4%
iperov2019-03-28 17:55:42 +04:00
689aefeb2f
Converter: fix for input images with ch > 3
iperov2019-03-28 11:12:19 +04:00
3cf3bb786e
rename samples to samplelib
iperov2019-03-27 10:44:13 +04:00
773e8d80e0
SampleGeneratorFace added ability to specify random seed (currently unused)
iperov2019-03-27 10:39:21 +04:00
24dd4ef000
get rid of random module
iperov2019-03-27 10:28:07 +04:00
b8efb4cbba
enhancing landmarks extractor by using s3fd second pass inside second pass, it will be x2 slower, but time will be saved due to more images will be marked properly works on 2GB+
iperov2019-03-18 10:25:24 +04:00
1d56585f33
fix ability for auto re-extract dst faces, after this update DFL torrent extractor scripts will not work.
iperov2019-03-18 01:12:04 +04:00
b03d62a29b
extractor: added ability to re-extract dst faces with transferring all necessary meta data
iperov2019-03-18 00:43:07 +04:00
00e56299c4
S3FD: wrong small face check fix
iperov2019-03-18 00:30:08 +04:00
f009edae79
FANSegmentator for developing
iperov2019-03-16 20:58:20 +04:00
1c8845d558
model fansegmentator for developing purposes
iperov2019-03-16 20:57:51 +04:00
f3b343c0e5
small fixed and refactorings
iperov2019-03-16 20:55:51 +04:00
d6a45763a2
SAE: added option "Use CA weights": Initialize network with 'Convolution Aware' weights. This may help to achieve a higher accuracy model, but consumes time at first run.
iperov2019-03-16 12:54:36 +04:00
58763756f5
SAE: removed simple_optimizer . Added optimizer mode for tensorflow only (NVIDIA cards), allows to train x2-x3 bigger networks with normal Adam optimizer, consuming VRAM and CPU power.
iperov2019-03-13 11:54:17 +04:00