github/DeepFaceLab
1
0
Fork 0
mirror of https://github.com/iperov/DeepFaceLab.git synced 2025-08-19 21:13:20 -07:00
Code Issues Projects Releases Packages Wiki Activity Actions

all models: removed options 'src_scale_mod', and 'sort samples by yaw as target'

Browse source 
If you want, you can manually remove unnecessary angles from src faceset after sort by yaw.

Optimized sample generators (CPU workers). Now they consume less amount of RAM and work faster.

added
4.2.other) data_src/dst util faceset pack.bat
	Packs /aligned/ samples into one /aligned/samples.pak file.
	After that, all faces will be deleted.

4.2.other) data_src/dst util faceset unpack.bat
	unpacks faces from /aligned/samples.pak to /aligned/ dir.
	After that, samples.pak will be deleted.

Packed faceset load and work faster.
This commit is contained in:
Colombo 2019-12-21 23:16:55 +04:00
commit 50f892d57d
26 changed files with 577 additions and 433 deletions
Download patch file Download diff file Expand all files Collapse all files

61
models/ModelBase.py
View file

@ -28,12 +28,10 @@ class ModelBase(object):
ask_write_preview_history=True,
ask_target_iter=True,
ask_batch_size=True,
ask_sort_by_yaw=True,
ask_random_flip=True,
ask_src_scale_mod=True):
ask_random_flip=True, **kwargs):
device_args['force_gpu_idx'] = device_args.get('force_gpu_idx',-1)
device_args['cpu_only'] = device_args.get('cpu_only',False)
device_args['cpu_only'] = True if debug else device_args.get('cpu_only',False)
if device_args['force_gpu_idx'] == -1 and not device_args['cpu_only']:
idxs_names_list = nnlib.device.getValidDevicesIdxsWithNamesList()
@ -115,13 +113,6 @@ class ModelBase(object):
else:
self.batch_size = self.options.get('batch_size', 0)
if ask_sort_by_yaw:
if (self.iter == 0 or ask_override):
default_sort_by_yaw = self.options.get('sort_by_yaw', False)
self.options['sort_by_yaw'] = io.input_bool("Feed faces to network sorted by yaw? (y/n ?:help skip:%s) : " % (yn_str[default_sort_by_yaw]), default_sort_by_yaw, help_message="NN will not learn src face directions that don't match dst face directions. Do not use if the dst face has hair that covers the jaw." )
else:
self.options['sort_by_yaw'] = self.options.get('sort_by_yaw', False)
if ask_random_flip:
default_random_flip = self.options.get('random_flip', True)
if (self.iter == 0 or ask_override):
@ -129,12 +120,6 @@ class ModelBase(object):
else:
self.options['random_flip'] = self.options.get('random_flip', default_random_flip)
if ask_src_scale_mod:
if (self.iter == 0):
self.options['src_scale_mod'] = np.clip( io.input_int("Src face scale modifier % ( -30...30, ?:help skip:0) : ", 0, help_message="If src face shape is wider than dst, try to decrease this value to get a better result."), -30, 30)
else:
self.options['src_scale_mod'] = self.options.get('src_scale_mod', 0)
self.autobackup = self.options.get('autobackup', False)
if not self.autobackup and 'autobackup' in self.options:
self.options.pop('autobackup')
@ -151,10 +136,6 @@ class ModelBase(object):
self.sort_by_yaw = self.options.get('sort_by_yaw',False)
self.random_flip = self.options.get('random_flip',True)
self.src_scale_mod = self.options.get('src_scale_mod',0)
if self.src_scale_mod == 0 and 'src_scale_mod' in self.options:
self.options.pop('src_scale_mod')
self.onInitializeOptions(self.iter == 0, ask_override)
nnlib.import_all(self.device_config)
@ -226,14 +207,14 @@ class ModelBase(object):
io.destroy_window(wnd_name)
else:
self.sample_for_preview = self.generate_next_sample()
try:
self.get_static_preview()
except:
self.sample_for_preview = self.generate_next_sample()
self.last_sample = self.sample_for_preview
###Generate text summary of model hyperparameters
#Find the longest key name and value string. Used as column widths.
width_name = max([len(k) for k in self.options.keys()] + [17]) + 1 # Single space buffer to left edge. Minimum of 17, the length of the longest static string used "Current iteration"
@ -414,48 +395,48 @@ class ModelBase(object):
def load_weights_safe(self, model_filename_list, optimizer_filename_list=[]):
exec(nnlib.code_import_all, locals(), globals())
loaded = []
not_loaded = []
for mf in model_filename_list:
model, filename = mf
filename = self.get_strpath_storage_for_file(filename)
if Path(filename).exists():
loaded += [ mf ]
if issubclass(model.__class__, keras.optimizers.Optimizer):
opt = model
try:
with open(filename, "rb") as f:
fd = pickle.loads(f.read())
weights = fd.get('weights', None)
if weights is not None:
opt.set_weights(weights)
except Exception as e:
print ("Unable to load ", filename)
else:
model.load_weights(filename)
else:
not_loaded += [ mf ]
return loaded, not_loaded
def save_weights_safe(self, model_filename_list):
exec(nnlib.code_import_all, locals(), globals())
for model, filename in model_filename_list:
filename = self.get_strpath_storage_for_file(filename) + '.tmp'
if issubclass(model.__class__, keras.optimizers.Optimizer):
opt = model
try:
try:
fd = {}
symbolic_weights = getattr(opt, 'weights')
if symbolic_weights:
@ -464,8 +445,8 @@ class ModelBase(object):
with open(filename, 'wb') as f:
f.write( pickle.dumps(fd) )
except Exception as e:
print ("Unable to save ", filename)
else:
print ("Unable to save ", filename)
else:
model.save_weights( filename)
rename_list = model_filename_list
@ -518,7 +499,7 @@ class ModelBase(object):
#overridable
def on_success_train_one_iter(self):
pass
def train_one_iter(self):
sample = self.generate_next_sample()
iter_time = time.time()
@ -546,7 +527,7 @@ class ModelBase(object):
cv2_imwrite (filepath, img )
self.on_success_train_one_iter()
self.iter += 1
return self.iter, iter_time

4
models/Model_AVATAR/Model.py
View file

@ -16,9 +16,7 @@ class AVATARModel(ModelBase):
def __init__(self, *args, **kwargs):
super().__init__(*args, **kwargs,
ask_sort_by_yaw=False,
ask_random_flip=False,
ask_src_scale_mod=False)
ask_random_flip=False)
#override
def onInitializeOptions(self, is_first_run, ask_override):

4
models/Model_DEV_FANSEG/Model.py
View file

@ -13,9 +13,7 @@ class Model(ModelBase):
ask_enable_autobackup=False,
ask_write_preview_history=False,
ask_target_iter=False,
ask_sort_by_yaw=False,
ask_random_flip=False,
ask_src_scale_mod=False)
ask_random_flip=False)
#override
def onInitializeOptions(self, is_first_run, ask_override):

12
models/Model_DEV_FUNIT/Model.py
View file

@ -16,9 +16,7 @@ class FUNITModel(ModelBase):
def __init__(self, *args, **kwargs):
super().__init__(*args, **kwargs,
ask_sort_by_yaw=False,
ask_random_flip=False,
ask_src_scale_mod=False)
ask_random_flip=False)
#override
def onInitializeOptions(self, is_first_run, ask_override):
@ -87,19 +85,19 @@ class FUNITModel(ModelBase):
self.set_training_data_generators ([
SampleGeneratorFacePerson(self.training_data_src_path, debug=self.is_debug(), batch_size=self.batch_size,
sample_process_options=SampleProcessor.Options(random_flip=True, rotation_range=[0,0] ),
output_sample_types=output_sample_types, person_id_mode=1, use_caching=True, generators_count=1 ),
output_sample_types=output_sample_types, person_id_mode=1, ),
SampleGeneratorFacePerson(self.training_data_src_path, debug=self.is_debug(), batch_size=self.batch_size,
sample_process_options=SampleProcessor.Options(random_flip=True, rotation_range=[0,0] ),
output_sample_types=output_sample_types, person_id_mode=1, use_caching=True, generators_count=1 ),
output_sample_types=output_sample_types, person_id_mode=1, ),
SampleGeneratorFacePerson(self.training_data_dst_path, debug=self.is_debug(), batch_size=self.batch_size,
sample_process_options=SampleProcessor.Options(random_flip=True, rotation_range=[0,0]),
output_sample_types=output_sample_types1, person_id_mode=1, use_caching=True, generators_count=1 ),
output_sample_types=output_sample_types1, person_id_mode=1, ),
SampleGeneratorFacePerson(self.training_data_dst_path, debug=self.is_debug(), batch_size=self.batch_size,
sample_process_options=SampleProcessor.Options(random_flip=True, rotation_range=[0,0]),
output_sample_types=output_sample_types1, person_id_mode=1, use_caching=True, generators_count=1 ),
output_sample_types=output_sample_types1, person_id_mode=1, ),
])
#override

4
models/Model_DEV_POSEEST/Model.py
View file

@ -15,9 +15,7 @@ class Model(ModelBase):
ask_enable_autobackup=False,
ask_write_preview_history=False,
ask_target_iter=False,
ask_sort_by_yaw=False,
ask_random_flip=False,
ask_src_scale_mod=False)
ask_random_flip=False)
#override
def onInitializeOptions(self, is_first_run, ask_override):

5
models/Model_DF/Model.py
View file

@ -50,9 +50,8 @@ class Model(ModelBase):
{ 'types': (t.IMG_TRANSFORMED, t.FACE_TYPE_FULL, t.MODE_M), 'resolution':128} ]
self.set_training_data_generators ([
SampleGeneratorFace(self.training_data_src_path, sort_by_yaw_target_samples_path=self.training_data_dst_path if self.sort_by_yaw else None,
debug=self.is_debug(), batch_size=self.batch_size,
sample_process_options=SampleProcessor.Options(random_flip=self.random_flip, scale_range=np.array([-0.05, 0.05])+self.src_scale_mod / 100.0 ),
SampleGeneratorFace(self.training_data_src_path, debug=self.is_debug(), batch_size=self.batch_size,
sample_process_options=SampleProcessor.Options(random_flip=self.random_flip, scale_range=np.array([-0.05, 0.05]) ),
output_sample_types=output_sample_types),
SampleGeneratorFace(self.training_data_dst_path, debug=self.is_debug(), batch_size=self.batch_size,

5
models/Model_H128/Model.py
View file

@ -60,9 +60,8 @@ class Model(ModelBase):
{ 'types': (t.IMG_TRANSFORMED, t.FACE_TYPE_HALF, t.MODE_M), 'resolution':128} ]
self.set_training_data_generators ([
SampleGeneratorFace(self.training_data_src_path, sort_by_yaw_target_samples_path=self.training_data_dst_path if self.sort_by_yaw else None,
debug=self.is_debug(), batch_size=self.batch_size,
sample_process_options=SampleProcessor.Options(random_flip=self.random_flip, scale_range=np.array([-0.05, 0.05])+self.src_scale_mod / 100.0 ),
SampleGeneratorFace(self.training_data_src_path, debug=self.is_debug(), batch_size=self.batch_size,
sample_process_options=SampleProcessor.Options(random_flip=self.random_flip, scale_range=np.array([-0.05, 0.05]) ),
output_sample_types=output_sample_types ),
SampleGeneratorFace(self.training_data_dst_path, debug=self.is_debug(), batch_size=self.batch_size,

5
models/Model_H64/Model.py
View file

@ -61,9 +61,8 @@ class Model(ModelBase):
{ 'types': (t.IMG_TRANSFORMED, t.FACE_TYPE_HALF, t.MODE_M), 'resolution':64} ]
self.set_training_data_generators ([
SampleGeneratorFace(self.training_data_src_path, sort_by_yaw_target_samples_path=self.training_data_dst_path if self.sort_by_yaw else None,
debug=self.is_debug(), batch_size=self.batch_size,
sample_process_options=SampleProcessor.Options(random_flip=self.random_flip, scale_range=np.array([-0.05, 0.05])+self.src_scale_mod / 100.0 ),
SampleGeneratorFace(self.training_data_src_path, debug=self.is_debug(), batch_size=self.batch_size,
sample_process_options=SampleProcessor.Options(random_flip=self.random_flip, scale_range=np.array([-0.05, 0.05]) ),
output_sample_types=output_sample_types),
SampleGeneratorFace(self.training_data_dst_path, debug=self.is_debug(), batch_size=self.batch_size,

5
models/Model_LIAEF128/Model.py
View file

@ -55,9 +55,8 @@ class Model(ModelBase):
{ 'types': (t.IMG_TRANSFORMED, t.FACE_TYPE_FULL, t.MODE_M), 'resolution':128} ]
self.set_training_data_generators ([
SampleGeneratorFace(self.training_data_src_path, sort_by_yaw_target_samples_path=self.training_data_dst_path if self.sort_by_yaw else None,
debug=self.is_debug(), batch_size=self.batch_size,
sample_process_options=SampleProcessor.Options(random_flip=self.random_flip, scale_range=np.array([-0.05, 0.05])+self.src_scale_mod / 100.0 ),
SampleGeneratorFace(self.training_data_src_path, debug=self.is_debug(), batch_size=self.batch_size,
sample_process_options=SampleProcessor.Options(random_flip=self.random_flip, scale_range=np.array([-0.05, 0.05]) ),
output_sample_types=output_sample_types),
SampleGeneratorFace(self.training_data_dst_path, debug=self.is_debug(), batch_size=self.batch_size,

8
models/Model_Quick96/Model.py
View file

@ -18,14 +18,12 @@ class Quick96Model(ModelBase):
ask_write_preview_history=False,
ask_target_iter=False,
ask_batch_size=False,
ask_sort_by_yaw=False,
ask_random_flip=False,
ask_src_scale_mod=False)
ask_random_flip=False)
#override
def onInitialize(self):
exec(nnlib.import_all(), locals(), globals())
self.set_vram_batch_requirements({1.5:2,2:4})#,3:4,4:8})
self.set_vram_batch_requirements({1.5:2,2:4})
resolution = self.resolution = 96
@ -171,7 +169,7 @@ class Quick96Model(ModelBase):
self.set_training_data_generators ([
SampleGeneratorFace(self.training_data_src_path, debug=self.is_debug(), batch_size=self.batch_size,
sample_process_options=SampleProcessor.Options(random_flip=False, scale_range=np.array([-0.05, 0.05])+self.src_scale_mod / 100.0 ),
sample_process_options=SampleProcessor.Options(random_flip=False, scale_range=np.array([-0.05, 0.05]) ),
output_sample_types = [ {'types' : (t.IMG_WARPED_TRANSFORMED, t.FACE_TYPE_FULL, t.MODE_BGR), 'resolution': resolution, 'normalize_tanh':True },
{'types' : (t.IMG_TRANSFORMED, t.FACE_TYPE_FULL, t.MODE_BGR), 'resolution': resolution, 'normalize_tanh':True },
{'types' : (t.IMG_TRANSFORMED, t.FACE_TYPE_FULL, t.MODE_M), 'resolution': resolution } ]

7
models/Model_SAE/Model.py
View file

@ -466,18 +466,15 @@ class SAEModel(ModelBase):
training_data_src_path = self.training_data_src_path
training_data_dst_path = self.training_data_dst_path
sort_by_yaw = self.sort_by_yaw
if self.pretrain and self.pretraining_data_path is not None:
training_data_src_path = self.pretraining_data_path
training_data_dst_path = self.pretraining_data_path
sort_by_yaw = False
self.set_training_data_generators ([
SampleGeneratorFace(training_data_src_path, sort_by_yaw_target_samples_path=training_data_dst_path if sort_by_yaw else None,
random_ct_samples_path=training_data_dst_path if self.options['ct_mode'] != 'none' else None,
SampleGeneratorFace(training_data_src_path, random_ct_samples_path=training_data_dst_path if self.options['ct_mode'] != 'none' else None,
debug=self.is_debug(), batch_size=self.batch_size,
sample_process_options=SampleProcessor.Options(random_flip=self.random_flip, scale_range=np.array([-0.05, 0.05])+self.src_scale_mod / 100.0 ),
sample_process_options=SampleProcessor.Options(random_flip=self.random_flip, scale_range=np.array([-0.05, 0.05]) ),
output_sample_types = [ {'types' : (t.IMG_WARPED_TRANSFORMED, face_type, t_mode_bgr), 'resolution':resolution, 'ct_mode': self.options['ct_mode'] },
{'types' : (t.IMG_TRANSFORMED, face_type, t_mode_bgr), 'resolution': resolution, 'ct_mode': self.options['ct_mode'] },
{'types' : (t.IMG_TRANSFORMED, face_type, t.MODE_M), 'resolution': resolution } ]

11
models/Model_SAEHD/Model.py
View file

@ -529,26 +529,23 @@ class SAEHDModel(ModelBase):
training_data_src_path = self.training_data_src_path
training_data_dst_path = self.training_data_dst_path
sort_by_yaw = self.sort_by_yaw
if self.pretrain and self.pretraining_data_path is not None:
training_data_src_path = self.pretraining_data_path
training_data_dst_path = self.pretraining_data_path
sort_by_yaw = False
t_img_warped = t.IMG_WARPED_TRANSFORMED if self.options['random_warp'] else t.IMG_TRANSFORMED
self.set_training_data_generators ([
SampleGeneratorFace(training_data_src_path, sort_by_yaw_target_samples_path=training_data_dst_path if sort_by_yaw else None,
random_ct_samples_path=training_data_dst_path if self.options['ct_mode'] != 'none' else None,
debug=self.is_debug(), batch_size=self.batch_size, use_caching=False,
sample_process_options=SampleProcessor.Options(random_flip=self.random_flip, scale_range=np.array([-0.05, 0.05])+self.src_scale_mod / 100.0 ),
SampleGeneratorFace(training_data_src_path, random_ct_samples_path=training_data_dst_path if self.options['ct_mode'] != 'none' else None,
debug=self.is_debug(), batch_size=self.batch_size,
sample_process_options=SampleProcessor.Options(random_flip=self.random_flip, scale_range=np.array([-0.05, 0.05]) ),
output_sample_types = [ {'types' : (t_img_warped, face_type, t_mode_bgr), 'resolution':resolution, 'ct_mode': self.options['ct_mode'] },
{'types' : (t.IMG_TRANSFORMED, face_type, t_mode_bgr), 'resolution': resolution, 'ct_mode': self.options['ct_mode'] },
{'types' : (t.IMG_TRANSFORMED, face_type, t.MODE_M), 'resolution': resolution } ]
),
SampleGeneratorFace(training_data_dst_path, debug=self.is_debug(), batch_size=self.batch_size, use_caching=False,
SampleGeneratorFace(training_data_dst_path, debug=self.is_debug(), batch_size=self.batch_size,
sample_process_options=SampleProcessor.Options(random_flip=self.random_flip, ),
output_sample_types = [ {'types' : (t_img_warped, face_type, t_mode_bgr), 'resolution':resolution},
{'types' : (t.IMG_TRANSFORMED, face_type, t_mode_bgr), 'resolution': resolution},