71f22957a6
SAEHD: random scale increased to -0.15+0.15. Improved lr_dropout capability to reach lower value of the loss.
...
When random_warp is off, inter_AB network is no longer trained to keep the face more src-like.
added option Random hue/saturation/light intensity applied to the src face set only at the input of the neural network. Stabilizes color perturbations during face swapping. Reduces the quality of the color transfer by selecting the closest one in the src faceset. Thus the src faceset must be diverse enough. Typical fine value is 0.05.
AMP: random scale increased to -0.15+0.15. Improved lr_dropout capability to reach lower value of the loss.
2021-10-18 11:03:23 +04:00
f71a9e6bbd
update SampleProcessor.py
2021-10-18 11:02:54 +04:00
a1ba64be10
_
2021-10-18 10:51:25 +04:00
14cc9d4e5f
_
2021-10-17 21:37:22 +04:00
d8c7cc3d93
DFLJPG increase max chunk size for xseg data
2021-10-13 14:35:26 +04:00
f64b2495d9
dev_misc: code to extract microsoft/FaceSynthetics dataset
2021-10-12 17:23:57 +04:00
f48e852de3
SAEHD: better training for bg_style_power
2021-10-12 15:17:41 +04:00
f700bd5954
update windows magnet url
2021-10-11 17:31:37 +04:00
80cae7d9f5
_
2021-10-11 15:03:34 +04:00
91f0f30a78
fix save in trainer
2021-10-11 15:02:41 +04:00
0a203bb25e
update windows magnet link
2021-10-09 14:38:00 +04:00
b1b5d6f482
AMP, SAEHD: In the sample generator, the random scaling was increased from -0.05+0.05 to -0.125+0.125 , which improves the generalization of faces.
2021-10-09 13:58:46 +04:00
f8469fe4d7
fix sample processor
2021-10-09 13:56:53 +04:00
3aa2b56eda
fix
2021-09-30 00:40:59 +04:00
c6d0c547b7
update optimizers
2021-09-29 22:06:46 +04:00
9e0079c6a0
SAEHD, AMP: removed the implicit function of periodically retraining last 16 “high-loss” samples
2021-09-29 16:48:54 +04:00
33ff0be722
_
2021-09-29 16:41:43 +04:00
8a897f236f
fix blur_out_mask for tf 1.13, fix export model for dx12 live build.
2021-09-11 00:00:15 +04:00
c1cee5d3ca
fix
2021-09-09 23:48:31 +04:00
7b96538cfb
upd readme
2021-09-09 23:20:02 +04:00
9f73ed0e7e
upd pic
2021-09-06 21:13:30 +04:00
cc133c0702
update mini tutorial link
2021-09-06 20:00:10 +04:00
b9c9552db6
upd windows magnet link
2021-09-06 17:20:51 +04:00
bd0b408a9b
fix
2021-09-06 16:11:23 +04:00
335da71db6
saehd: remove sample processor count limitation
2021-09-06 14:16:19 +04:00
70c9ee9463
change face_style_power
2021-09-06 14:11:08 +04:00
d301a4799f
_
2021-09-06 14:07:20 +04:00
b1990d421a
AMP: don't train inters with random_warp-off
2021-09-03 13:50:31 +04:00
3fe8ce86b1
AMP: enable inter_src+inter_dst training
2021-08-30 18:52:29 +04:00
f346e35e7f
fix
2021-08-30 12:57:19 +04:00
6e094d873d
SAEHD: removed 'dst_denoise' option. Added -t arhi option.
2021-08-29 11:48:55 +04:00
01f1a084b4
AMP: remove dst_denoise option
2021-08-29 11:46:51 +04:00
e53be5e22d
fix
2021-08-24 13:18:51 +04:00
91187ecb95
AMP: some arhi change results training stabilization. New options blur_out_mask + rtm_dst_denoise. Sample processors count are no more limited to 8, thus if you have AMD processor with 16+ cores, increase paging file size.
2021-08-24 12:58:23 +04:00
8e63666390
SAEHD: add 'Blur out mask' and 'Denoise DST faceset' options.
2021-08-20 17:07:44 +04:00
e41f87e682
SampleProcessor: add denoise filter
2021-08-20 17:06:43 +04:00
b256b07e03
fix gaussian_blur
2021-08-20 17:06:11 +04:00
57f3393ec2
tf add unused TanhPolar
2021-08-20 17:05:26 +04:00
56e70edc46
fix error in model saving
2021-08-19 23:18:04 +04:00
26c83f6e35
upd readme
2021-08-15 09:23:29 +04:00
664c8fd105
update related works
2021-08-13 16:59:18 +04:00
8a455fc238
upd magnet link
2021-08-13 13:23:28 +04:00
fce90a1b44
update magnet link
2021-08-12 22:33:31 +04:00
f95b3cdd33
upd deepfake_progress.png
2021-08-12 20:03:24 +04:00
60ade74327
update deepfake_progress.png
2021-08-12 20:01:08 +04:00
f6bd45cce0
upd deepfake_progress.png
2021-08-12 19:29:09 +04:00
c5584fbda0
XSeg: changed pretrain mode
2021-08-12 17:01:38 +04:00
623eb3856d
upd README.md
2021-08-12 16:59:09 +04:00
4aa33d3ba8
_
2021-08-12 14:37:57 +04:00
a299324166
_
2021-08-12 14:28:36 +04:00
iperov