mirror of
https://github.com/iperov/DeepFaceLab.git
synced 2025-07-07 13:32:09 -07:00
AVATAR fixes
This commit is contained in:
iperov
parent
612ef5155e
commit
4aaac5e42e
1 changed files with 120 additions and 86 deletions
|
|
@ -12,7 +12,7 @@ class Model(ModelBase):
|
|||
decoder64_srcH5 = 'decoder64_src.h5'
|
||||
decoder64_dstH5 = 'decoder64_dst.h5'
|
||||
encoder256H5 = 'encoder256.h5'
|
||||
decoder256_srcH5 = 'decoder256_src.h5'
|
||||
decoder256H5 = 'decoder256.h5'
|
||||
|
||||
#override
|
||||
def onInitialize(self, **in_options):
|
||||
|
|
@ -20,71 +20,79 @@ class Model(ModelBase):
|
|||
keras = self.keras
|
||||
K = keras.backend
|
||||
|
||||
self.set_vram_batch_requirements( {4:8,5:16,6:20,7:24,8:32,9:48} )
|
||||
|
||||
self.encoder64, self.decoder64_src, self.decoder64_dst, self.encoder256, self.decoder256_src = self.BuildAE()
|
||||
img_shape64 = (64,64,1)
|
||||
img_shape256 = (256,256,3)
|
||||
|
||||
self.set_vram_batch_requirements( {3.5:8,4:8,5:12,6:16,7:24,8:32,9:48} )
|
||||
if self.batch_size < 4:
|
||||
self.batch_size = 4
|
||||
|
||||
img_shape64, img_shape256, self.encoder64, self.decoder64_src, self.decoder64_dst, self.encoder256, self.decoder256 = self.Build()
|
||||
|
||||
if not self.is_first_run():
|
||||
self.encoder64.load_weights (self.get_strpath_storage_for_file(self.encoder64H5))
|
||||
self.decoder64_src.load_weights (self.get_strpath_storage_for_file(self.decoder64_srcH5))
|
||||
self.decoder64_dst.load_weights (self.get_strpath_storage_for_file(self.decoder64_dstH5))
|
||||
self.encoder256.load_weights (self.get_strpath_storage_for_file(self.encoder256H5))
|
||||
self.decoder256_src.load_weights (self.get_strpath_storage_for_file(self.decoder256_srcH5))
|
||||
self.decoder256.load_weights (self.get_strpath_storage_for_file(self.decoder256H5))
|
||||
|
||||
if self.is_training_mode:
|
||||
self.encoder64, self.decoder64_src, self.decoder64_dst, self.encoder256, self.decoder256_src = self.to_multi_gpu_model_if_possible ( [self.encoder64, self.decoder64_src, self.decoder64_dst, self.encoder256, self.decoder256_src] )
|
||||
self.encoder64, self.decoder64_src, self.decoder64_dst, self.encoder256, self.decoder256 = self.to_multi_gpu_model_if_possible ( [self.encoder64, self.decoder64_src, self.decoder64_dst, self.encoder256, self.decoder256] )
|
||||
|
||||
input_src_64 = keras.layers.Input(img_shape64)
|
||||
input_src_target64 = keras.layers.Input(img_shape64)
|
||||
input_src_target256 = keras.layers.Input(img_shape256)
|
||||
input_dst_64 = keras.layers.Input(img_shape64)
|
||||
input_dst_target64 = keras.layers.Input(img_shape64)
|
||||
input_A_warped64 = keras.layers.Input(img_shape64)
|
||||
input_A_target64 = keras.layers.Input(img_shape64)
|
||||
input_B_warped64 = keras.layers.Input(img_shape64)
|
||||
input_B_target64 = keras.layers.Input(img_shape64)
|
||||
|
||||
src_code64 = self.encoder64(input_src_64)
|
||||
dst_code64 = self.encoder64(input_dst_64)
|
||||
A_code64 = self.encoder64(input_A_warped64)
|
||||
B_code64 = self.encoder64(input_B_warped64)
|
||||
|
||||
rec_src64 = self.decoder64_src(src_code64)
|
||||
rec_dst64 = self.decoder64_dst(dst_code64)
|
||||
A_rec64 = self.decoder64_src(A_code64)
|
||||
B_rec64 = self.decoder64_dst(B_code64)
|
||||
|
||||
src64_loss = tf_dssim(tf, input_src_target64, rec_src64)
|
||||
dst64_loss = tf_dssim(tf, input_dst_target64, rec_dst64)
|
||||
total64_loss = src64_loss + dst64_loss
|
||||
A64_loss = tf_dssim(tf, input_A_target64, A_rec64)
|
||||
B64_loss = tf_dssim(tf, input_B_target64, B_rec64)
|
||||
total64_loss = A64_loss + B64_loss
|
||||
|
||||
self.ed64_train = K.function ([input_src_64, input_src_target64, input_dst_64, input_dst_target64],[K.mean(total64_loss)],
|
||||
self.ed64_train = K.function ([input_A_warped64, input_A_target64, input_B_warped64, input_B_target64],[K.mean(total64_loss)],
|
||||
self.keras.optimizers.Adam(lr=5e-5, beta_1=0.5, beta_2=0.999).get_updates(total64_loss, self.encoder64.trainable_weights + self.decoder64_src.trainable_weights + self.decoder64_dst.trainable_weights)
|
||||
)
|
||||
|
||||
src_code256 = self.encoder256(input_src_64)
|
||||
rec_src256 = self.decoder256_src(src_code256)
|
||||
src256_loss = tf_dssim(tf, input_src_target256, rec_src256)
|
||||
|
||||
self.ed256_train = K.function ([input_src_64, input_src_target256],[K.mean(src256_loss)],
|
||||
self.keras.optimizers.Adam(lr=5e-5, beta_1=0.5, beta_2=0.999).get_updates(src256_loss, self.encoder256.trainable_weights + self.decoder256_src.trainable_weights)
|
||||
)
|
||||
|
||||
src_code256 = self.encoder256(rec_src64)
|
||||
rec_src256 = self.decoder256_src(src_code256)
|
||||
self.A64_view = K.function ([input_A_warped64], [A_rec64])
|
||||
self.B64_view = K.function ([input_B_warped64], [B_rec64])
|
||||
|
||||
input_A_warped64 = keras.layers.Input(img_shape64)
|
||||
input_A_target256 = keras.layers.Input(img_shape256)
|
||||
A_code256 = self.encoder256(input_A_warped64)
|
||||
A_rec256 = self.decoder256(A_code256)
|
||||
|
||||
self.src256_view = K.function ([input_src_64], [rec_src256])
|
||||
input_B_warped64 = keras.layers.Input(img_shape64)
|
||||
B_code64 = self.encoder64(input_B_warped64)
|
||||
BA_rec64 = self.decoder64_src(B_code64)
|
||||
BA_code256 = self.encoder256(BA_rec64)
|
||||
BA_rec256 = self.decoder256(BA_code256)
|
||||
|
||||
total256_loss = K.mean( tf_dssim(tf, input_A_target256, A_rec256) )
|
||||
|
||||
self.ed256_train = K.function ([input_A_warped64, input_A_target256],[total256_loss],
|
||||
self.keras.optimizers.Adam(lr=5e-5, beta_1=0.5, beta_2=0.999).get_updates(total256_loss, self.encoder256.trainable_weights + self.decoder256.trainable_weights)
|
||||
)
|
||||
|
||||
self.A256_view = K.function ([input_A_warped64], [A_rec256])
|
||||
self.BA256_view = K.function ([input_B_warped64], [BA_rec256])
|
||||
|
||||
if self.is_training_mode:
|
||||
from models import TrainingDataGenerator
|
||||
f = TrainingDataGenerator.SampleTypeFlags
|
||||
self.set_training_data_generators ([
|
||||
TrainingDataGenerator(TrainingDataType.FACE, self.training_data_src_path, debug=self.is_debug(), batch_size=self.batch_size, output_sample_types=[
|
||||
[f.WARPED_TRANSFORMED | f.HALF_FACE | f.MODE_G, 64],
|
||||
[f.TRANSFORMED | f.HALF_FACE | f.MODE_G, 64],
|
||||
[f.WARPED_TRANSFORMED | f.HALF_FACE | f.MODE_BGR, 64],
|
||||
[f.TRANSFORMED | f.HALF_FACE | f.MODE_BGR, 64],
|
||||
[f.TRANSFORMED | f.FULL_FACE | f.MODE_BGR, 256],
|
||||
[f.SOURCE | f.HALF_FACE | f.MODE_G, 64],
|
||||
[f.SOURCE | f.HALF_FACE | f.MODE_GGG, 256] ] ),
|
||||
[f.SOURCE | f.HALF_FACE | f.MODE_BGR, 64],
|
||||
[f.SOURCE | f.HALF_FACE | f.MODE_BGR, 256] ] ),
|
||||
|
||||
TrainingDataGenerator(TrainingDataType.FACE, self.training_data_dst_path, debug=self.is_debug(), batch_size=self.batch_size, output_sample_types=[
|
||||
[f.WARPED_TRANSFORMED | f.HALF_FACE | f.MODE_G, 64],
|
||||
[f.TRANSFORMED | f.HALF_FACE | f.MODE_G, 64],
|
||||
[f.SOURCE | f.HALF_FACE | f.MODE_G, 64],
|
||||
[f.SOURCE | f.HALF_FACE | f.MODE_GGG, 256] ] )
|
||||
[f.WARPED_TRANSFORMED | f.HALF_FACE | f.MODE_BGR, 64],
|
||||
[f.TRANSFORMED | f.HALF_FACE | f.MODE_BGR, 64],
|
||||
[f.SOURCE | f.HALF_FACE | f.MODE_BGR, 64],
|
||||
[f.SOURCE | f.HALF_FACE | f.MODE_BGR, 256] ] )
|
||||
])
|
||||
#override
|
||||
def onSave(self):
|
||||
|
|
@ -92,73 +100,100 @@ class Model(ModelBase):
|
|||
[self.decoder64_src, self.get_strpath_storage_for_file(self.decoder64_srcH5)],
|
||||
[self.decoder64_dst, self.get_strpath_storage_for_file(self.decoder64_dstH5)],
|
||||
[self.encoder256, self.get_strpath_storage_for_file(self.encoder256H5)],
|
||||
[self.decoder256_src, self.get_strpath_storage_for_file(self.decoder256_srcH5)],
|
||||
[self.decoder256, self.get_strpath_storage_for_file(self.decoder256H5)],
|
||||
] )
|
||||
|
||||
#override
|
||||
def onTrainOneEpoch(self, sample):
|
||||
warped_src64, target_src64, target_src256, target_src_source64_G, target_src_source256_GGG = sample[0]
|
||||
warped_dst64, target_dst64, target_dst_source64_G, target_dst_source256_GGG = sample[1]
|
||||
warped_src64, target_src64, target_src256, target_src_source64, target_src_source256 = sample[0]
|
||||
warped_dst64, target_dst64, target_dst_source64, target_dst_source256 = sample[1]
|
||||
|
||||
loss64, = self.ed64_train ([warped_src64, target_src64, warped_dst64, target_dst64])
|
||||
loss256, = self.ed256_train ([warped_src64, target_src256])
|
||||
|
||||
return ( ('loss64', loss64), ('loss256', loss256) )
|
||||
return ( ('loss64', loss64), ('loss256', loss256), )
|
||||
|
||||
#override
|
||||
def onGetPreview(self, sample):
|
||||
n_samples = 4
|
||||
test_B = sample[1][2][0:n_samples]
|
||||
test_B256 = sample[1][3][0:n_samples]
|
||||
sample_src64_source = sample[0][3][0:4]
|
||||
sample_src256_source = sample[0][4][0:4]
|
||||
|
||||
BB, = self.src256_view ([test_B])
|
||||
sample_dst64_source = sample[1][2][0:4]
|
||||
sample_dst256_source = sample[1][3][0:4]
|
||||
|
||||
SRC64, = self.A64_view ([sample_src64_source])
|
||||
DST64, = self.B64_view ([sample_dst64_source])
|
||||
SRCDST64, = self.A64_view ([sample_dst64_source])
|
||||
DSTSRC64, = self.B64_view ([sample_src64_source])
|
||||
|
||||
SRC_x1_256, = self.A256_view ([sample_src64_source])
|
||||
DST_x2_256, = self.BA256_view ([sample_dst64_source])
|
||||
|
||||
b1 = np.concatenate ( (
|
||||
np.concatenate ( (sample_src64_source[0], SRC64[0], sample_src64_source[1], SRC64[1], ), axis=1),
|
||||
np.concatenate ( (sample_src64_source[1], SRC64[1], sample_src64_source[3], SRC64[3], ), axis=1),
|
||||
np.concatenate ( (sample_dst64_source[0], DST64[0], sample_dst64_source[1], DST64[1], ), axis=1),
|
||||
np.concatenate ( (sample_dst64_source[2], DST64[2], sample_dst64_source[3], DST64[3], ), axis=1),
|
||||
), axis=0 )
|
||||
|
||||
b2 = np.concatenate ( (
|
||||
np.concatenate ( (sample_src64_source[0], DSTSRC64[0], sample_src64_source[1], DSTSRC64[1], ), axis=1),
|
||||
np.concatenate ( (sample_src64_source[2], DSTSRC64[2], sample_src64_source[3], DSTSRC64[3], ), axis=1),
|
||||
np.concatenate ( (sample_dst64_source[0], SRCDST64[0], sample_dst64_source[1], SRCDST64[1], ), axis=1),
|
||||
np.concatenate ( (sample_dst64_source[2], SRCDST64[2], sample_dst64_source[3], SRCDST64[3], ), axis=1),
|
||||
|
||||
), axis=0 )
|
||||
|
||||
result = np.concatenate ( ( np.concatenate ( (b1, sample_src256_source[0], SRC_x1_256[0] ), axis=1 ),
|
||||
np.concatenate ( (b2, sample_dst256_source[0], DST_x2_256[0] ), axis=1 ),
|
||||
), axis = 0 )
|
||||
|
||||
st = []
|
||||
for i in range(n_samples // 2):
|
||||
st.append ( np.concatenate ( (
|
||||
test_B256[i*2+0], BB[i*2+0], test_B256[i*2+1], BB[i*2+1],
|
||||
), axis=1) )
|
||||
return [ ('AVATAR', np.concatenate ( st, axis=0 ) ) ]
|
||||
return [ ('AVATAR', result ) ]
|
||||
|
||||
def predictor_func (self, img):
|
||||
x, = self.src256_view ([ np.expand_dims(img, 0) ])[0]
|
||||
x, = self.BA256_view ([ np.expand_dims(img, 0) ])[0]
|
||||
return x
|
||||
|
||||
#override
|
||||
def get_converter(self, **in_options):
|
||||
return ConverterAvatar(self.predictor_func, predictor_input_size=64, output_size=256, **in_options)
|
||||
|
||||
def BuildAE(self):
|
||||
def Build(self):
|
||||
keras, K = self.keras, self.keras.backend
|
||||
|
||||
img_shape64 = (64,64,3)
|
||||
img_shape256 = (256,256,3)
|
||||
|
||||
def Encoder(_input):
|
||||
x = keras.layers.convolutional.Conv2D(90, kernel_size=5, strides=1, padding='same')(_input)
|
||||
x = keras.layers.convolutional.Conv2D(90, kernel_size=5, strides=1, padding='same')(x)
|
||||
x = keras.layers.MaxPooling2D(pool_size=(3, 3), strides=2, padding='same')(x)
|
||||
x = _input
|
||||
x = self.keras.layers.convolutional.Conv2D(90, kernel_size=5, strides=1, padding='same')(x)
|
||||
x = self.keras.layers.convolutional.Conv2D(90, kernel_size=5, strides=1, padding='same')(x)
|
||||
x = self.keras.layers.MaxPooling2D(pool_size=(3, 3), strides=2, padding='same')(x)
|
||||
|
||||
x = keras.layers.convolutional.Conv2D(180, kernel_size=3, strides=1, padding='same')(x)
|
||||
x = keras.layers.convolutional.Conv2D(180, kernel_size=3, strides=1, padding='same')(x)
|
||||
x = keras.layers.MaxPooling2D(pool_size=(3, 3), strides=2, padding='same')(x)
|
||||
x = self.keras.layers.convolutional.Conv2D(180, kernel_size=3, strides=1, padding='same')(x)
|
||||
x = self.keras.layers.convolutional.Conv2D(180, kernel_size=3, strides=1, padding='same')(x)
|
||||
x = self.keras.layers.MaxPooling2D(pool_size=(3, 3), strides=2, padding='same')(x)
|
||||
|
||||
x = keras.layers.convolutional.Conv2D(360, kernel_size=3, strides=1, padding='same')(x)
|
||||
x = keras.layers.convolutional.Conv2D(360, kernel_size=3, strides=1, padding='same')(x)
|
||||
x = keras.layers.MaxPooling2D(pool_size=(3, 3), strides=2, padding='same')(x)
|
||||
x = self.keras.layers.convolutional.Conv2D(360, kernel_size=3, strides=1, padding='same')(x)
|
||||
x = self.keras.layers.convolutional.Conv2D(360, kernel_size=3, strides=1, padding='same')(x)
|
||||
x = self.keras.layers.MaxPooling2D(pool_size=(3, 3), strides=2, padding='same')(x)
|
||||
|
||||
x = keras.layers.Dense (1024)(x)
|
||||
x = keras.layers.advanced_activations.LeakyReLU(0.1)(x)
|
||||
x = keras.layers.Dropout(0.5)(x)
|
||||
x = self.keras.layers.Dense (1024)(x)
|
||||
x = self.keras.layers.advanced_activations.LeakyReLU(0.1)(x)
|
||||
x = self.keras.layers.Dropout(0.5)(x)
|
||||
|
||||
x = self.keras.layers.Dense (1024)(x)
|
||||
x = self.keras.layers.advanced_activations.LeakyReLU(0.1)(x)
|
||||
x = self.keras.layers.Dropout(0.5)(x)
|
||||
x = self.keras.layers.Flatten()(x)
|
||||
x = self.keras.layers.Dense (64)(x)
|
||||
|
||||
x = keras.layers.Dense (1024)(x)
|
||||
x = keras.layers.advanced_activations.LeakyReLU(0.1)(x)
|
||||
x = keras.layers.Dropout(0.5)(x)
|
||||
x = keras.layers.Flatten()(x)
|
||||
x = keras.layers.Dense (64)(x)
|
||||
return keras.models.Model (_input, x)
|
||||
|
||||
encoder256 = Encoder( keras.layers.Input ( (64, 64, 1) ) )
|
||||
encoder64 = Encoder( keras.layers.Input ( (64, 64, 1) ) )
|
||||
encoder256 = Encoder( keras.layers.Input (img_shape64) )
|
||||
encoder64 = Encoder( keras.layers.Input (img_shape64) )
|
||||
|
||||
def decoder256_3(encoder):
|
||||
def decoder256(encoder):
|
||||
decoder_input = keras.layers.Input ( K.int_shape(encoder.outputs[0])[1:] )
|
||||
x = decoder_input
|
||||
x = self.keras.layers.Dense(16 * 16 * 720)(x)
|
||||
|
|
@ -170,18 +205,18 @@ class Model(ModelBase):
|
|||
x = keras.layers.convolutional.Conv2D(3, kernel_size=5, padding='same', activation='sigmoid')(x)
|
||||
return keras.models.Model(decoder_input, x)
|
||||
|
||||
def decoder64_1(encoder):
|
||||
def decoder64(encoder):
|
||||
decoder_input = keras.layers.Input ( K.int_shape(encoder.outputs[0])[1:] )
|
||||
x = decoder_input
|
||||
x = self.keras.layers.Dense(8 * 8 * 720)(x)
|
||||
x = keras.layers.Reshape ( (8,8,720) )(x)
|
||||
x = keras.layers.Reshape ( (8, 8, 720) )(x)
|
||||
x = upscale(keras, x, 360)
|
||||
x = upscale(keras, x, 180)
|
||||
x = upscale(keras, x, 90)
|
||||
x = keras.layers.convolutional.Conv2D(1, kernel_size=5, padding='same', activation='sigmoid')(x)
|
||||
x = keras.layers.convolutional.Conv2D(3, kernel_size=5, padding='same', activation='sigmoid')(x)
|
||||
return keras.models.Model(decoder_input, x)
|
||||
|
||||
return encoder64, decoder64_1(encoder64), decoder64_1(encoder64), encoder256, decoder256_3(encoder256)
|
||||
return img_shape64, img_shape256, encoder64, decoder64(encoder64), decoder64(encoder64), encoder256, decoder256(encoder256)
|
||||
|
||||
from models import ConverterBase
|
||||
from facelib import FaceType
|
||||
|
|
@ -205,7 +240,7 @@ class ConverterAvatar(ConverterBase):
|
|||
|
||||
#override
|
||||
def dummy_predict(self):
|
||||
self.predictor ( np.zeros ( (self.predictor_input_size, self.predictor_input_size,1), dtype=np.float32) )
|
||||
self.predictor ( np.zeros ( (self.predictor_input_size, self.predictor_input_size,3), dtype=np.float32) )
|
||||
|
||||
#override
|
||||
def convert_image (self, img_bgr, img_face_landmarks, debug):
|
||||
|
|
@ -213,9 +248,8 @@ class ConverterAvatar(ConverterBase):
|
|||
|
||||
face_mat = LandmarksProcessor.get_transform_mat (img_face_landmarks, self.predictor_input_size, face_type=FaceType.HALF )
|
||||
predictor_input_bgr = cv2.warpAffine( img_bgr, face_mat, (self.predictor_input_size, self.predictor_input_size), flags=cv2.INTER_LANCZOS4 )
|
||||
predictor_input_g = np.expand_dims(cv2.cvtColor(predictor_input_bgr, cv2.COLOR_BGR2GRAY),-1)
|
||||
|
||||
predicted_bgr = self.predictor ( predictor_input_g )
|
||||
|
||||
predicted_bgr = self.predictor ( predictor_input_bgr )
|
||||
|
||||
output = cv2.resize ( predicted_bgr, (self.output_size, self.output_size), cv2.INTER_LANCZOS4 )
|
||||
if debug:
|
||||
|
|
|
|||
Loading…
Add table
Add a link
Reference in a new issue