github/DeepFaceLab
1
0
Fork 0
mirror of https://github.com/iperov/DeepFaceLab.git synced 2025-07-07 05:22:06 -07:00
Code Issues Projects Releases Packages Wiki Activity Actions
1391 commits 3 branches 2 tags 866 MiB
Commit graph

53 commits

Author SHA1 Message Date
leno3003
5999e9f20c added sumbmodules 2022-03-31 11:44:00 +02:00
leno3003
d59009d4b9 added sumbmodules 2022-03-31 11:29:51 +02:00
leno3003
c8df5d3f18 added sumbmodules 2022-03-31 11:29:13 +02:00
leno3003
c66db83b17 added sumbmodules 2022-03-31 11:27:31 +02:00
leno3003
2023661008 added sumbmodules 2022-03-31 10:59:45 +02:00
leno3003
ab7fe2e42d added sumbmodules 2022-03-31 10:57:12 +02:00
leno3003
a23a778507 added sumbmodules 2022-03-30 18:12:11 +02:00
leno3003
9410e20984 added sumbmodules 2022-03-30 18:05:26 +02:00
leno3003
b7c5e649e3 added sumbmodules 2022-03-30 17:49:16 +02:00
leno3003
c2622c95da added sumbmodules 2022-03-30 16:17:55 +02:00
leno3003
31811f46f5 added sumbmodules 2022-03-30 16:16:55 +02:00
leno3003
39aa9f91ba added sumbmodules 2022-03-30 16:15:18 +02:00
leno3003
578c0ea158 added sumbmodules 2022-03-30 16:13:09 +02:00
leno3003
d71759d7a0 added sumbmodules 2022-03-30 15:54:43 +02:00
leno3003
a3de74223f added sumbmodules 2022-03-30 15:53:17 +02:00
leno3003
c40d7ee52f added sumbmodules 2022-03-30 15:28:08 +02:00
leno3003
a0ecfc063b added sumbmodules 2022-03-30 14:33:37 +02:00
leno3003
d415f56135 added sumbmodules 2022-03-30 14:32:47 +02:00
leno3003
25f93f313f added sumbmodules 2022-03-30 14:30:08 +02:00
leno3003
a900a1f02f added sumbmodules 2022-03-30 14:15:01 +02:00
leno3003
6d68692e74 added sumbmodules 2022-03-30 14:10:30 +02:00
leno3003
d4279606f4 added sumbmodules 2022-03-30 14:04:21 +02:00
leno3003
887286edc5 added sumbmodules 2022-03-30 12:49:01 +02:00
leno3003
dc39e72fb9 added sumbmodules 2022-03-30 12:33:35 +02:00
leno3003
dee168e5a6 automatized arguments with pickle 2022-03-28 22:43:24 +02:00
leno3003
afd87df67a automatized arguments with pickle 2022-03-28 18:30:15 +02:00
leno3003
21024c64ad automatized arguments with pickle 2022-03-28 17:03:36 +02:00
leno3003
688b3e1567 automatized arguments with pickle 2022-03-28 15:54:06 +02:00
Colombo
3283dce96a interact: change to waitKeyEx 2020-07-03 18:43:18 +04:00
Colombo
ddb6bcf416 fix 2020-06-24 11:44:24 +04:00
Colombo
2b8e8f0554 . 2020-06-20 22:05:30 +04:00
Colombo
0c2e1c3944 SAEHD: 
Maximum resolution is increased to 640.

‘hd’ archi is removed. ‘hd’ was experimental archi created to remove subpixel shake, but ‘lr_dropout’ and ‘disable random warping’ do that better.

‘uhd’ is renamed to ‘-u’
dfuhd and liaeuhd will be automatically renamed to df-u and liae-u in existing models.

Added new experimental archi (key -d) which doubles the resolution using the same computation cost.
It is mean same configs will be x2 faster, or for example you can set 448 resolution and it will train as 224.
Strongly recommended not to train from scratch and use pretrained models.

New archi naming:
'df' keeps more identity-preserved face.
'liae' can fix overly different face shapes.
'-u' increased likeness of the face.
'-d' (experimental) doubling the resolution using the same computation cost
Examples: df, liae, df-d, df-ud, liae-ud, ...

Improved GAN training (GAN_power option).  It was used for dst model, but actually we don’t need it for dst.
Instead, a second src GAN model with x2 smaller patch size was added, so the overall quality for hi-res models should be higher.

Added option ‘Uniform yaw distribution of samples (y/n)’:
	Helps to fix blurry side faces due to small amount of them in the faceset.

Quick96:
	Now based on df-ud archi and 20% faster.

XSeg trainer:
	Improved sample generator.
Now it randomly adds the background from other samples.
Result is reduced chance of random mask noise on the area outside the face.
Now you can specify ‘batch_size’ in range 2-16.

Reduced size of samples with applied XSeg mask. Thus size of packed samples with applied xseg mask is also reduced.
 2020-06-19 09:45:55 +04:00
Colombo
2fe86faf01 fix Win10 input dialog 2020-04-02 10:27:39 +04:00
Colombo
61472cdaf7 global refactoring and fixes, 
removed support of extracted(aligned) PNG faces. Use old builds to convert from PNG to JPG.

fanseg model file in facelib/ is renamed
 2020-03-13 08:09:00 +04:00
Colombo
cbff72f597 _ 2020-02-29 15:53:48 +04:00
Colombo
6d122fd742 _ 2020-02-29 15:50:17 +04:00
Colombo
516e23584e _ 2020-02-29 15:48:31 +04:00
Colombo
803f7f9717 _ 2020-02-29 15:40:11 +04:00
Colombo
b093112f2f _ 2020-02-29 15:39:43 +04:00
Colombo
a21c55cf25 _ 2020-02-29 15:36:37 +04:00
Colombo
f55a65c304 _ 2020-02-29 15:34:27 +04:00
Colombo
55162d6e6c _ 2020-02-29 15:32:36 +04:00
Colombo
27e24d5d28 _ 2020-02-29 15:24:14 +04:00
Colombo
63c0bdbf16 )_ 2020-02-29 15:10:59 +04:00
Colombo
98d3657cbc _ 2020-02-29 15:00:46 +04:00
Colombo
7f609d0f0a fix for linux 2020-02-29 14:49:46 +04:00
Colombo
dbf28eeebf fix for linux 2020-02-29 14:36:10 +04:00
Colombo
5eca4958b7 fix for linux 2020-02-29 14:34:08 +04:00
Colombo
90ae68ae61 fix input_in_time for linux 2020-02-29 11:05:04 +04:00
Colombo
76ca79216e Upgraded to TF version 1.13.2 
Removed the wait at first launch for most graphics cards.

Increased speed of training by 10-20%, but you have to retrain all models from scratch.

SAEHD:

added option 'use float16'
	Experimental option. Reduces the model size by half.
	Increases the speed of training.
	Decreases the accuracy of the model.
	The model may collapse or not train.
	Model may not learn the mask in large resolutions.

true_face_training option is replaced by
"True face power". 0.0000 .. 1.0
Experimental option. Discriminates the result face to be more like the src face. Higher value - stronger discrimination.
Comparison - https://i.imgur.com/czScS9q.png
 2020-01-25 21:58:19 +04:00