mirror of
https://github.com/iperov/DeepFaceLab.git
synced 2025-07-06 13:02:15 -07:00
fix ModelBase, nnlib
This commit is contained in:
iperov
parent
90a7d4b1e7
commit
a9026ccb67
2 changed files with 27 additions and 67 deletions
|
|
@ -23,8 +23,9 @@ class ModelBase(object):
|
|||
def __init__(self, model_path, training_data_src_path=None, training_data_dst_path=None, debug = False, device_args = None):
|
||||
|
||||
device_args['force_gpu_idx'] = device_args.get('force_gpu_idx',-1)
|
||||
device_args['cpu_only'] = device_args.get('cpu_only',False)
|
||||
|
||||
if device_args['force_gpu_idx'] == -1:
|
||||
if device_args['force_gpu_idx'] == -1 and not device_args['cpu_only']:
|
||||
idxs_names_list = nnlib.device.getValidDevicesIdxsWithNamesList()
|
||||
if len(idxs_names_list) > 1:
|
||||
io.log_info ("You have multi GPUs in a system: ")
|
||||
|
|
@ -34,10 +35,7 @@ class ModelBase(object):
|
|||
device_args['force_gpu_idx'] = io.input_int("Which GPU idx to choose? ( skip: best GPU ) : ", -1, [ x[0] for x in idxs_names_list] )
|
||||
self.device_args = device_args
|
||||
|
||||
nnlib.import_all ( nnlib.DeviceConfig(allow_growth=False, **self.device_args) )
|
||||
self.device_config = nnlib.active_DeviceConfig
|
||||
self.keras = nnlib.keras
|
||||
self.K = nnlib.keras.backend
|
||||
self.device_config = nnlib.DeviceConfig(allow_growth=False, **self.device_args)
|
||||
|
||||
io.log_info ("Loading model...")
|
||||
|
||||
|
|
@ -127,8 +125,12 @@ class ModelBase(object):
|
|||
if self.src_scale_mod == 0:
|
||||
self.options.pop('src_scale_mod')
|
||||
|
||||
|
||||
self.onInitializeOptions(self.iter == 0, ask_override)
|
||||
|
||||
nnlib.import_all(self.device_config)
|
||||
self.keras = nnlib.keras
|
||||
self.K = nnlib.keras.backend
|
||||
|
||||
self.onInitialize()
|
||||
|
||||
self.options['batch_size'] = self.batch_size
|
||||
|
|
|
|||
|
|
@ -148,18 +148,12 @@ NLayerDiscriminator = nnlib.NLayerDiscriminator
|
|||
suppressor.__exit__()
|
||||
|
||||
@staticmethod
|
||||
def import_keras(device_config = None):
|
||||
def import_keras(device_config):
|
||||
if nnlib.keras is not None:
|
||||
return nnlib.code_import_keras
|
||||
|
||||
if device_config is None:
|
||||
device_config = nnlib.active_DeviceConfig
|
||||
|
||||
nnlib.active_DeviceConfig = device_config
|
||||
|
||||
if "tensorflow" in device_config.backend:
|
||||
nnlib._import_tf(device_config)
|
||||
device_config = nnlib.active_DeviceConfig
|
||||
elif device_config.backend == "plaidML":
|
||||
os.environ["KERAS_BACKEND"] = "plaidml.keras.backend"
|
||||
os.environ["PLAIDML_DEVICE_IDS"] = ",".join ( [ nnlib.device.getDeviceID(idx) for idx in device_config.gpu_idxs] )
|
||||
|
|
@ -435,38 +429,16 @@ NLayerDiscriminator = nnlib.NLayerDiscriminator
|
|||
nnlib.Scale = Scale
|
||||
|
||||
class AdamCPU(keras.optimizers.Optimizer):
|
||||
"""Adam optimizer.
|
||||
Default parameters follow those provided in the original paper.
|
||||
# Arguments
|
||||
lr: float >= 0. Learning rate.
|
||||
beta_1: float, 0 < beta < 1. Generally close to 1.
|
||||
beta_2: float, 0 < beta < 1. Generally close to 1.
|
||||
epsilon: float >= 0. Fuzz factor. If `None`, defaults to `K.epsilon()`.
|
||||
decay: float >= 0. Learning rate decay over each update.
|
||||
amsgrad: boolean. Whether to apply the AMSGrad variant of this
|
||||
algorithm from the paper "On the Convergence of Adam and
|
||||
Beyond".
|
||||
# References
|
||||
- [Adam - A Method for Stochastic Optimization](
|
||||
https://arxiv.org/abs/1412.6980v8)
|
||||
- [On the Convergence of Adam and Beyond](
|
||||
https://openreview.net/forum?id=ryQu7f-RZ)
|
||||
"""
|
||||
|
||||
def __init__(self, lr=0.001, beta_1=0.9, beta_2=0.999,
|
||||
epsilon=None, decay=0., amsgrad=False, tf_cpu_mode=0, **kwargs):
|
||||
tf_cpu_mode=0, **kwargs):
|
||||
super(AdamCPU, self).__init__(**kwargs)
|
||||
with K.name_scope(self.__class__.__name__):
|
||||
self.iterations = K.variable(0, dtype='int64', name='iterations')
|
||||
self.lr = K.variable(lr, name='lr')
|
||||
self.beta_1 = K.variable(beta_1, name='beta_1')
|
||||
self.beta_2 = K.variable(beta_2, name='beta_2')
|
||||
self.decay = K.variable(decay, name='decay')
|
||||
if epsilon is None:
|
||||
epsilon = K.epsilon()
|
||||
self.epsilon = epsilon
|
||||
self.initial_decay = decay
|
||||
self.amsgrad = amsgrad
|
||||
|
||||
self.epsilon = K.epsilon()
|
||||
self.tf_cpu_mode = tf_cpu_mode
|
||||
|
||||
@keras.legacy.interfaces.legacy_get_updates_support
|
||||
|
|
@ -474,12 +446,8 @@ NLayerDiscriminator = nnlib.NLayerDiscriminator
|
|||
grads = self.get_gradients(loss, params)
|
||||
self.updates = [K.update_add(self.iterations, 1)]
|
||||
|
||||
|
||||
lr = self.lr
|
||||
if self.initial_decay > 0:
|
||||
lr = lr * (1. / (1. + self.decay * K.cast(self.iterations,
|
||||
K.dtype(self.decay))))
|
||||
|
||||
|
||||
t = K.cast(self.iterations, K.floatx()) + 1
|
||||
lr_t = lr * (K.sqrt(1. - K.pow(self.beta_2, t)) /
|
||||
(1. - K.pow(self.beta_1, t)))
|
||||
|
|
@ -488,21 +456,13 @@ NLayerDiscriminator = nnlib.NLayerDiscriminator
|
|||
with K.tf.device("/cpu:0"):
|
||||
ms = [K.zeros(K.int_shape(p), dtype=K.dtype(p)) for p in params]
|
||||
vs = [K.zeros(K.int_shape(p), dtype=K.dtype(p)) for p in params]
|
||||
if self.amsgrad:
|
||||
vhats = [K.zeros(K.int_shape(p), dtype=K.dtype(p)) for p in params]
|
||||
else:
|
||||
vhats = [K.zeros(1) for _ in params]
|
||||
else:
|
||||
ms = [K.zeros(K.int_shape(p), dtype=K.dtype(p)) for p in params]
|
||||
vs = [K.zeros(K.int_shape(p), dtype=K.dtype(p)) for p in params]
|
||||
if self.amsgrad:
|
||||
vhats = [K.zeros(K.int_shape(p), dtype=K.dtype(p)) for p in params]
|
||||
else:
|
||||
vhats = [K.zeros(1) for _ in params]
|
||||
|
||||
self.weights = [self.iterations] + ms + vs + vhats
|
||||
self.weights = [self.iterations] + ms + vs
|
||||
|
||||
for p, g, m, v, vhat in zip(params, grads, ms, vs, vhats):
|
||||
for p, g, m, v in zip(params, grads, ms, vs):
|
||||
if self.tf_cpu_mode == 2:
|
||||
with K.tf.device("/cpu:0"):
|
||||
m_t = (self.beta_1 * m) + (1. - self.beta_1) * g
|
||||
|
|
@ -511,12 +471,7 @@ NLayerDiscriminator = nnlib.NLayerDiscriminator
|
|||
m_t = (self.beta_1 * m) + (1. - self.beta_1) * g
|
||||
v_t = (self.beta_2 * v) + (1. - self.beta_2) * K.square(g)
|
||||
|
||||
if self.amsgrad:
|
||||
vhat_t = K.maximum(vhat, v_t)
|
||||
p_t = p - lr_t * m_t / (K.sqrt(vhat_t) + self.epsilon)
|
||||
self.updates.append(K.update(vhat, vhat_t))
|
||||
else:
|
||||
p_t = p - lr_t * m_t / (K.sqrt(v_t) + self.epsilon)
|
||||
p_t = p - lr_t * m_t / (K.sqrt(v_t) + self.epsilon)
|
||||
|
||||
self.updates.append(K.update(m, m_t))
|
||||
self.updates.append(K.update(v, v_t))
|
||||
|
|
@ -532,10 +487,8 @@ NLayerDiscriminator = nnlib.NLayerDiscriminator
|
|||
def get_config(self):
|
||||
config = {'lr': float(K.get_value(self.lr)),
|
||||
'beta_1': float(K.get_value(self.beta_1)),
|
||||
'beta_2': float(K.get_value(self.beta_2)),
|
||||
'decay': float(K.get_value(self.decay)),
|
||||
'epsilon': self.epsilon,
|
||||
'amsgrad': self.amsgrad}
|
||||
'beta_2': float(K.get_value(self.beta_2))
|
||||
}
|
||||
base_config = super(AdamCPU, self).get_config()
|
||||
return dict(list(base_config.items()) + list(config.items()))
|
||||
|
||||
|
|
@ -561,7 +514,7 @@ NLayerDiscriminator = nnlib.NLayerDiscriminator
|
|||
|
||||
|
||||
@staticmethod
|
||||
def import_keras_contrib(device_config = None):
|
||||
def import_keras_contrib(device_config):
|
||||
if nnlib.keras_contrib is not None:
|
||||
return nnlib.code_import_keras_contrib
|
||||
|
||||
|
|
@ -588,7 +541,12 @@ NLayerDiscriminator = nnlib.NLayerDiscriminator
|
|||
|
||||
@staticmethod
|
||||
def import_all(device_config = None):
|
||||
if nnlib.code_import_all is None:
|
||||
if nnlib.code_import_all is None:
|
||||
if device_config is None:
|
||||
device_config = nnlib.active_DeviceConfig
|
||||
else:
|
||||
nnlib.active_DeviceConfig = device_config
|
||||
|
||||
nnlib.import_keras(device_config)
|
||||
nnlib.import_keras_contrib(device_config)
|
||||
nnlib.code_import_all = compile (nnlib.code_import_keras_string + '\n'
|
||||
|
|
@ -600,8 +558,8 @@ NLayerDiscriminator = nnlib.NLayerDiscriminator
|
|||
|
||||
@staticmethod
|
||||
def __initialize_all_functions():
|
||||
exec (nnlib.import_keras(), locals(), globals())
|
||||
exec (nnlib.import_keras_contrib(), locals(), globals())
|
||||
exec (nnlib.import_keras(nnlib.active_DeviceConfig), locals(), globals())
|
||||
exec (nnlib.import_keras_contrib(nnlib.active_DeviceConfig), locals(), globals())
|
||||
|
||||
class DSSIMMSEMaskLoss(object):
|
||||
def __init__(self, mask, is_mse=False):
|
||||
|
|
|
|||
Loading…
Add table
Add a link
Reference in a new issue