github/DeepFaceLab
1
0
Fork 0
mirror of https://github.com/iperov/DeepFaceLab.git synced 2025-08-21 22:13:20 -07:00
Code Issues Projects Releases Packages Wiki Activity Actions

Merge branch 'master' into feature/ms-ssim-loss-2

Browse source 
# Conflicts:
#	models/Model_SAEHD/Model.py
This commit is contained in:
jh 2021-03-24 06:15:44 -07:00
commit c428ca0b6c
7 changed files with 97 additions and 29 deletions
Download patch file Download diff file Expand all files Collapse all files

12
CHANGELOG.md
View file

@ -10,6 +10,14 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
- [Freezeable layers (encoder/decoder/etc.)](https://github.com/faceshiftlabs/DeepFaceLab/tree/feature/freezable-weights)
- [GAN stability improvements](https://github.com/faceshiftlabs/DeepFaceLab/tree/feature/gan-updates)
## [1.3.0] - 2020-03-20
### Added
- [Background Power training option](doc/features/background-power/README.md)
## [1.2.1] - 2020-03-20
### Fixed
- Fixes bug with `fs-aug` color mode.
## [1.2.0] - 2020-03-17
### Added
- [Random color training option](doc/features/random-color/README.md)
@ -45,7 +53,9 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
- Reset stale master branch to [seranus/DeepFaceLab](https://github.com/seranus/DeepFaceLab),
21 commits ahead of [iperov/DeepFaceLab](https://github.com/iperov/DeepFaceLab) ([compare](https://github.com/iperov/DeepFaceLab/compare/4818183...seranus:3f5ae05))
[Unreleased]: https://github.com/olivierlacan/keep-a-changelog/compare/v1.2.0...HEAD
[Unreleased]: https://github.com/olivierlacan/keep-a-changelog/compare/v1.3.0...HEAD
[1.3.0]: https://github.com/faceshiftlabs/DeepFaceLab/compare/v1.2.1...v1.3.0
[1.2.1]: https://github.com/faceshiftlabs/DeepFaceLab/compare/v1.2.0...v1.2.1
[1.2.0]: https://github.com/faceshiftlabs/DeepFaceLab/compare/v1.1.5...v1.2.0
[1.1.5]: https://github.com/faceshiftlabs/DeepFaceLab/compare/v1.1.4...v1.1.5
[1.1.4]: https://github.com/faceshiftlabs/DeepFaceLab/compare/v1.1.3...v1.1.4

24
core/imagelib/color_transfer.py
View file

@ -1,7 +1,7 @@
import cv2
import numpy as np
from numpy import linalg as npla
from random import random, shuffle, choice
import random
from scipy.stats import special_ortho_group
import scipy as sp
@ -371,12 +371,12 @@ def color_transfer(ct_mode, img_src, img_trg):
# imported from faceswap
def color_augmentation(img):
def color_augmentation(img, seed=None):
""" Color adjust RGB image """
face = img
face = np.clip(face*255.0, 0, 255).astype(np.uint8)
face = random_clahe(face)
face = random_lab(face)
face = random_clahe(face, seed)
face = random_lab(face, seed)
img[:, :, :3] = face
return (face / 255.0).astype(np.float32)
@ -400,13 +400,14 @@ def random_lab_rotation(image, seed=None):
return image
def random_lab(image):
def random_lab(image, seed=None):
""" Perform random color/lightness adjustment in L*a*b* colorspace """
random.seed(seed)
amount_l = 30 / 100
amount_ab = 8 / 100
randoms = [(random() * amount_l * 2) - amount_l, # L adjust
(random() * amount_ab * 2) - amount_ab, # A adjust
(random() * amount_ab * 2) - amount_ab] # B adjust
randoms = [(random.random() * amount_l * 2) - amount_l, # L adjust
(random.random() * amount_ab * 2) - amount_ab, # A adjust
(random.random() * amount_ab * 2) - amount_ab] # B adjust
image = cv2.cvtColor( # pylint:disable=no-member
image, cv2.COLOR_BGR2LAB).astype("float32") / 255.0 # pylint:disable=no-member
@ -419,15 +420,16 @@ def random_lab(image):
cv2.COLOR_LAB2BGR) # pylint:disable=no-member
return image
def random_clahe(image):
def random_clahe(image, seed=None):
""" Randomly perform Contrast Limited Adaptive Histogram Equalization """
contrast_random = random()
random.seed(seed)
contrast_random = random.random()
if contrast_random > 50 / 100:
return image
# base_contrast = image.shape[0] // 128
base_contrast = 1 # testing because it breaks on small sizes
grid_base = random() * 4
grid_base = random.random() * 4
contrast_adjustment = int(grid_base * (base_contrast / 2))
grid_size = base_contrast + contrast_adjustment

32
doc/features/background-power/README.md Normal file
View file

@ -0,0 +1,32 @@
# Background Power option
Allows you to train the model to include the background, which may help with areas around the mask.
Unlike **Background Style Power**, this does not use any additional VRAM, and does not require lowering the batch size.
- [DESCRIPTION](#description)
- [USAGE](#usage)
- [DIFFERENCE WITH BACKGROUND STYLE POWER](#difference-with-background-style-power)
*Examples trained with background power `0.3`:*
![](example.jpeg)![](example2.jpeg)
## DESCRIPTION
Applies the same loss calculation used for the area *inside* the mask, to the area *outside* the mask, multiplied with
the chosen background power value.
E.g. (simplified): Source Loss = Masked area image difference + Background Power * Non-masked area image difference
## USAGE
`[0.0] Background power ( 0.0..1.0 ?:help ) : 0.3`
## DIFFERENCE WITH BACKGROUND STYLE POWER
**Background Style Power** applies a loss to the source by comparing the background of the dest to that of the
predicted src/dest (5th column). This operation requires additional VRAM, due to the face that the predicted src/dest
outputs are not normally used in training (other then being viewable in the preview window).
**Background Power** does *not* use the src/dest images whatsoever, instead comparing the background of the predicted
source to that of the original source, and the same for the background of the dest images.

BIN
doc/features/background-power/example.jpeg Normal file
View file

Binary file not shown.
Side by side

After

Width:  |  Height:  |  Size: 129 KiB

BIN
doc/features/background-power/example2.jpeg Normal file
View file

Binary file not shown.
Side by side

After

Width:  |  Height:  |  Size: 121 KiB

24
models/Model_SAEHD/Model.py
View file

@ -56,6 +56,7 @@ class SAEHDModel(ModelBase):
default_ms_ssim_loss = self.options['ms_ssim_loss'] = self.load_or_def_option('ms_ssim_loss', False)
default_random_warp = self.options['random_warp'] = self.load_or_def_option('random_warp', True)
default_background_power = self.options['background_power'] = self.load_or_def_option('background_power', 0.0)
default_true_face_power = self.options['true_face_power'] = self.load_or_def_option('true_face_power', 0.0)
default_face_style_power = self.options['face_style_power'] = self.load_or_def_option('face_style_power', 0.0)
default_bg_style_power = self.options['bg_style_power'] = self.load_or_def_option('bg_style_power', 0.0)
@ -168,6 +169,8 @@ Examples: df, liae, df-d, df-ud, liae-ud, ...
else:
self.options['true_face_power'] = 0.0
self.options['background_power'] = np.clip ( io.input_number("Background power", default_background_power, add_info="0.0..1.0", help_message="Learn the area outside of the mask. Helps smooth out area near the mask boundaries. Can be used at any time"), 0.0, 1.0 )
self.options['face_style_power'] = np.clip ( io.input_number("Face style power", default_face_style_power, add_info="0.0..100.0", help_message="Learn the color of the predicted face to be the same as dst inside mask. If you want to use this option with 'whole_face' you have to use XSeg trained mask. Warning: Enable it only after 10k iters, when predicted face is clear enough to start learn style. Start from 0.001 value and check history changes. Enabling this option increases the chance of model collapse."), 0.0, 100.0 )
self.options['bg_style_power'] = np.clip ( io.input_number("Background style power", default_bg_style_power, add_info="0.0..100.0", help_message="Learn the area outside mask of the predicted face to be the same as dst. If you want to use this option with 'whole_face' you have to use XSeg trained mask. For whole_face you have to use XSeg trained mask. This can make face more like dst. Enabling this option increases the chance of model collapse. Typical value is 2.0"), 0.0, 100.0 )
@ -417,12 +420,15 @@ Examples: df, liae, df-d, df-ud, liae-ud, ...
gpu_target_dst_style_anti_masked = gpu_target_dst*(1.0 - gpu_target_dstm_style_blur)
gpu_target_src_anti_masked = gpu_target_src*(1.0-gpu_target_srcm_blur)
gpu_target_dst_anti_masked = gpu_target_dst_style_anti_masked
gpu_target_src_masked_opt = gpu_target_src*gpu_target_srcm_blur if masked_training else gpu_target_src
gpu_target_dst_masked_opt = gpu_target_dst_masked if masked_training else gpu_target_dst
gpu_pred_src_src_masked_opt = gpu_pred_src_src*gpu_target_srcm_blur if masked_training else gpu_pred_src_src
gpu_pred_src_src_anti_masked = gpu_pred_src_src*(1.0-gpu_target_srcm_blur)
gpu_pred_dst_dst_masked_opt = gpu_pred_dst_dst*gpu_target_dstm_blur if masked_training else gpu_pred_dst_dst
gpu_pred_dst_dst_anti_masked = gpu_pred_dst_dst*(1.0-gpu_target_dstm_blur)
gpu_psd_target_dst_style_masked = gpu_pred_src_dst*gpu_target_dstm_style_blur
gpu_psd_target_dst_style_anti_masked = gpu_pred_src_dst*(1.0 - gpu_target_dstm_style_blur)
@ -455,6 +461,15 @@ Examples: df, liae, df-d, df-ud, liae-ud, ...
else:
gpu_src_loss += tf.reduce_mean ( 10*tf.square( gpu_target_srcm - gpu_pred_src_srcm ),axis=[1,2,3] )
if self.options['background_power'] > 0:
bg_factor = self.options['background_power']
if resolution < 256:
gpu_src_loss += bg_factor * tf.reduce_mean ( 10*nn.dssim(gpu_target_src_anti_masked, gpu_pred_src_src_anti_masked, max_val=1.0, filter_size=int(resolution/11.6)), axis=[1])
else:
gpu_src_loss += bg_factor * tf.reduce_mean ( 5*nn.dssim(gpu_target_src_anti_masked, gpu_pred_src_src_anti_masked, max_val=1.0, filter_size=int(resolution/11.6)), axis=[1])
gpu_src_loss += bg_factor * tf.reduce_mean ( 5*nn.dssim(gpu_target_src_anti_masked, gpu_pred_src_src_anti_masked, max_val=1.0, filter_size=int(resolution/23.2)), axis=[1])
gpu_src_loss += bg_factor * tf.reduce_mean ( 10*tf.square ( gpu_target_src_anti_masked - gpu_pred_src_src_anti_masked ), axis=[1,2,3])
face_style_power = self.options['face_style_power'] / 100.0
if face_style_power != 0 and not self.pretrain:
gpu_src_loss += nn.style_loss(gpu_psd_target_dst_style_masked, gpu_target_dst_style_masked, gaussian_blur_radius=resolution//16, loss_weight=10000*face_style_power)
@ -488,6 +503,15 @@ Examples: df, liae, df-d, df-ud, liae-ud, ...
else:
gpu_dst_loss += tf.reduce_mean ( 300*tf.abs ( gpu_target_dst*gpu_target_part_mask - gpu_pred_dst_dst*gpu_target_part_mask ), axis=[1,2,3])
if self.options['background_power'] > 0:
bg_factor = self.options['background_power']
if resolution < 256:
gpu_dst_loss += bg_factor * tf.reduce_mean ( 10*nn.dssim(gpu_target_dst_anti_masked, gpu_pred_dst_dst_anti_masked, max_val=1.0, filter_size=int(resolution/11.6)), axis=[1])
else:
gpu_dst_loss += bg_factor * tf.reduce_mean ( 5*nn.dssim(gpu_target_dst_anti_masked, gpu_pred_dst_dst_anti_masked, max_val=1.0, filter_size=int(resolution/11.6)), axis=[1])
gpu_dst_loss += bg_factor * tf.reduce_mean ( 5*nn.dssim(gpu_target_dst_anti_masked, gpu_pred_dst_dst_anti_masked, max_val=1.0, filter_size=int(resolution/23.2)), axis=[1])
gpu_dst_loss += bg_factor * tf.reduce_mean ( 10*tf.square ( gpu_target_dst_anti_masked - gpu_pred_dst_dst_anti_masked ), axis=[1,2,3])
if self.options['ms_ssim_loss']:
gpu_dst_loss += 10 * nn.MsSsim(resolution)(gpu_target_dstm, gpu_pred_dst_dstm, max_val=1.0)
else:

2
samplelib/SampleProcessor.py
View file

@ -207,7 +207,7 @@ class SampleProcessor(object):
# Apply random color transfer
if ct_mode is not None and ct_sample is not None or ct_mode == 'fs-aug':
if ct_mode == 'fs-aug':
img = imagelib.color_augmentation(img)
img = imagelib.color_augmentation(img, sample_rnd_seed)
else:
if ct_sample_bgr is None:
ct_sample_bgr = ct_sample.load_bgr()