mirror of
https://github.com/iperov/DeepFaceLab.git
synced 2025-07-06 04:52:13 -07:00
optimized face sample generator, CPU load is significantly reduced
SAEHD: added new option GAN power 0.0 .. 10.0 Train the network in Generative Adversarial manner. Forces the neural network to learn small details of the face. You can enable/disable this option at any time, but better to enable it when the network is trained enough. Typical value is 1.0 GAN power with pretrain mode will not work. Example of enabling GAN on 81k iters +5k iters https://i.imgur.com/OdXHLhU.jpg https://i.imgur.com/CYAJmJx.jpg dfhd: default Decoder dimensions are now 48 the preview for 256 res is now correctly displayed fixed model naming/renaming/removing Improvements for those involved in post-processing in AfterEffects: Codec is reverted back to x264 in order to properly use in AfterEffects and video players. Merger now always outputs the mask to workspace\data_dst\merged_mask removed raw modes except raw-rgb raw-rgb mode now outputs selected face mask_mode (before square mask) 'export alpha mask' button is replaced by 'show alpha mask'. You can view the alpha mask without recompute the frames. 8) 'merged *.bat' now also output 'result_mask.' video file. 8) 'merged lossless' now uses x264 lossless codec (before PNG codec) result_mask video file is always lossless. Thus you can use result_mask video file as mask layer in the AfterEffects.
This commit is contained in:
Colombo
parent
80f285067a
commit
7386a9d6fd
28 changed files with 455 additions and 363 deletions
|
|
@ -14,7 +14,7 @@ class QModel(ModelBase):
|
|||
#override
|
||||
def on_initialize(self):
|
||||
device_config = nn.getCurrentDeviceConfig()
|
||||
self.model_data_format = "NCHW" if len(device_config.devices) != 0 else "NHWC"
|
||||
self.model_data_format = "NCHW" if len(device_config.devices) != 0 and not self.is_debug() else "NHWC"
|
||||
nn.initialize(data_format=self.model_data_format)
|
||||
tf = nn.tf
|
||||
|
||||
|
|
@ -167,9 +167,9 @@ class QModel(ModelBase):
|
|||
models_opt_device = '/GPU:0' if models_opt_on_gpu and self.is_training else '/CPU:0'
|
||||
optimizer_vars_on_cpu = models_opt_device=='/CPU:0'
|
||||
|
||||
input_nc = 3
|
||||
output_nc = 3
|
||||
bgr_shape = nn.get4Dshape(resolution,resolution,input_nc)
|
||||
input_ch = 3
|
||||
output_ch = 3
|
||||
bgr_shape = nn.get4Dshape(resolution,resolution,input_ch)
|
||||
mask_shape = nn.get4Dshape(resolution,resolution,1)
|
||||
lowest_dense_res = resolution // 16
|
||||
|
||||
|
|
@ -189,7 +189,7 @@ class QModel(ModelBase):
|
|||
|
||||
# Initializing model classes
|
||||
with tf.device (models_opt_device):
|
||||
self.encoder = Encoder(in_ch=input_nc, e_ch=e_dims, name='encoder')
|
||||
self.encoder = Encoder(in_ch=input_ch, e_ch=e_dims, name='encoder')
|
||||
encoder_out_ch = self.encoder.compute_output_channels ( (nn.tf_floatx, bgr_shape))
|
||||
|
||||
self.inter = Inter (in_ch=encoder_out_ch, lowest_dense_res=lowest_dense_res, ae_ch=ae_dims, ae_out_ch=ae_dims, d_ch=d_dims, name='inter')
|
||||
|
|
@ -228,7 +228,7 @@ class QModel(ModelBase):
|
|||
gpu_src_losses = []
|
||||
gpu_dst_losses = []
|
||||
gpu_src_dst_loss_gvs = []
|
||||
|
||||
|
||||
for gpu_id in range(gpu_count):
|
||||
with tf.device( f'/GPU:{gpu_id}' if len(devices) != 0 else f'/CPU:0' ):
|
||||
batch_slice = slice( gpu_id*bs_per_gpu, (gpu_id+1)*bs_per_gpu )
|
||||
|
|
@ -262,7 +262,7 @@ class QModel(ModelBase):
|
|||
gpu_target_dst_masked = gpu_target_dst*gpu_target_dstm_blur
|
||||
gpu_target_dst_anti_masked = gpu_target_dst*(1.0 - gpu_target_dstm_blur)
|
||||
|
||||
gpu_target_srcmasked_opt = gpu_target_src*gpu_target_srcm_blur if masked_training else gpu_target_src
|
||||
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
|
||||
|
|
@ -271,8 +271,8 @@ class QModel(ModelBase):
|
|||
gpu_psd_target_dst_masked = gpu_pred_src_dst*gpu_target_dstm_blur
|
||||
gpu_psd_target_dst_anti_masked = gpu_pred_src_dst*(1.0 - gpu_target_dstm_blur)
|
||||
|
||||
gpu_src_loss = tf.reduce_mean ( 10*nn.tf_dssim(gpu_target_srcmasked_opt, gpu_pred_src_src_masked_opt, max_val=1.0, filter_size=int(resolution/11.6)), axis=[1])
|
||||
gpu_src_loss += tf.reduce_mean ( 10*tf.square ( gpu_target_srcmasked_opt - gpu_pred_src_src_masked_opt ), axis=[1,2,3])
|
||||
gpu_src_loss = tf.reduce_mean ( 10*nn.tf_dssim(gpu_target_src_masked_opt, gpu_pred_src_src_masked_opt, max_val=1.0, filter_size=int(resolution/11.6)), axis=[1])
|
||||
gpu_src_loss += tf.reduce_mean ( 10*tf.square ( gpu_target_src_masked_opt - gpu_pred_src_src_masked_opt ), axis=[1,2,3])
|
||||
gpu_src_loss += tf.reduce_mean ( 10*tf.square( gpu_target_srcm - gpu_pred_src_srcm ),axis=[1,2,3] )
|
||||
|
||||
gpu_dst_loss = tf.reduce_mean ( 10*nn.tf_dssim(gpu_target_dst_masked_opt, gpu_pred_dst_dst_masked_opt, max_val=1.0, filter_size=int(resolution/11.6) ), axis=[1])
|
||||
|
|
@ -282,8 +282,8 @@ class QModel(ModelBase):
|
|||
gpu_src_losses += [gpu_src_loss]
|
||||
gpu_dst_losses += [gpu_dst_loss]
|
||||
|
||||
gpu_src_dst_loss = gpu_src_loss + gpu_dst_loss
|
||||
gpu_src_dst_loss_gvs += [ nn.tf_gradients ( gpu_src_dst_loss, self.src_dst_trainable_weights ) ]
|
||||
gpu_G_loss = gpu_src_loss + gpu_dst_loss
|
||||
gpu_src_dst_loss_gvs += [ nn.tf_gradients ( gpu_G_loss, self.src_dst_trainable_weights ) ]
|
||||
|
||||
|
||||
# Average losses and gradients, and create optimizer update ops
|
||||
|
|
@ -362,10 +362,9 @@ class QModel(ModelBase):
|
|||
training_data_src_path = self.training_data_src_path if not self.pretrain else self.get_pretraining_data_path()
|
||||
training_data_dst_path = self.training_data_dst_path if not self.pretrain else self.get_pretraining_data_path()
|
||||
|
||||
cpu_count = multiprocessing.cpu_count()
|
||||
|
||||
cpu_count = min(multiprocessing.cpu_count(), 8)
|
||||
src_generators_count = cpu_count // 2
|
||||
dst_generators_count = cpu_count - src_generators_count
|
||||
dst_generators_count = cpu_count // 2
|
||||
|
||||
self.set_training_data_generators ([
|
||||
SampleGeneratorFace(training_data_src_path, debug=self.is_debug(), batch_size=self.get_batch_size(),
|
||||
|
|
@ -396,18 +395,19 @@ class QModel(ModelBase):
|
|||
|
||||
#override
|
||||
def onTrainOneIter(self):
|
||||
|
||||
if self.get_iter() % 3 == 0 and self.last_samples is not None:
|
||||
( (warped_src, target_src, target_srcm), \
|
||||
(warped_dst, target_dst, target_dstm) ) = self.last_samples
|
||||
src_loss, dst_loss = self.src_dst_train (target_src, target_src, target_srcm,
|
||||
target_dst, target_dst, target_dstm)
|
||||
warped_src = target_src
|
||||
warped_dst = target_dst
|
||||
else:
|
||||
samples = self.last_samples = self.generate_next_samples()
|
||||
( (warped_src, target_src, target_srcm), \
|
||||
(warped_dst, target_dst, target_dstm) ) = samples
|
||||
|
||||
src_loss, dst_loss = self.src_dst_train (warped_src, target_src, target_srcm,
|
||||
warped_dst, target_dst, target_dstm)
|
||||
src_loss, dst_loss = self.src_dst_train (warped_src, target_src, target_srcm,
|
||||
warped_dst, target_dst, target_dstm)
|
||||
|
||||
return ( ('src_loss', src_loss), ('dst_loss', dst_loss), )
|
||||
|
||||
|
|
@ -440,8 +440,7 @@ class QModel(ModelBase):
|
|||
return result
|
||||
|
||||
def predictor_func (self, face=None):
|
||||
face = face[None,...]
|
||||
face = nn.to_data_format(face, self.model_data_format, "NHWC")
|
||||
face = nn.to_data_format(face[None,...], self.model_data_format, "NHWC")
|
||||
|
||||
bgr, mask_dst_dstm, mask_src_dstm = [ nn.to_data_format(x, "NHWC", self.model_data_format).astype(np.float32) for x in self.AE_merge (face) ]
|
||||
mask = mask_dst_dstm[0] * mask_src_dstm[0]
|
||||
|
|
|
|||
Loading…
Add table
Add a link
Reference in a new issue