def get_activation(fn_name,only_layer=False):
"""
Args:
fn_name ():
Returns:
Examples:
>>> print(get_activation('swish'))
"""
if fn_name is None:
return None
fn_modules = ['trident.layers.pytorch_activations', 'trident.backend.pytorch_ops', 'torch.nn.functional']
trident_fn_modules = ['trident.layers.pytorch_activations', 'trident.backend.pytorch_ops']
if only_layer:
fn_modules = ['trident.layers.pytorch_activations']
trident_fn_modules = ['trident.layers.pytorch_activations']
try:
if isinstance(fn_name, str):
if not only_layer and (camel2snake(fn_name)== fn_name or fn_name.lower()== fn_name):
if fn_name == 'p_relu' or fn_name == 'prelu':
return PRelu()
activation_fn = get_function(fn_name, trident_fn_modules)
return activation_fn
else:
try:
activation_fn = get_class(snake2camel(fn_name), fn_modules)
return activation_fn()
except Exception:
activation_fn = get_class(fn_name, fn_modules)
return activation_fn()
elif getattr(fn_name, '__module__', None) == 'trident.layers.pytorch_activations':
if inspect.isfunction(fn_name):
return partial(fn_name)
elif inspect.isclass(fn_name) and fn_name.__class__.__name__=="type":
return fn_name()
elif isinstance(fn_name, Layer):
return fn_name
elif inspect.isfunction(fn_name) and getattr(fn_name, '__module__', None) == 'trident.backend.pytorch_ops':
if only_layer:
activation_layer = get_class(snake2camel(fn_name.__name__), trident_fn_modules)
return activation_layer()
else:
return fn_name
else:
if callable(fn_name):
result = inspect.getfullargspec(fn_name)
if 1 <= len(result.args) <= 2:
return fn_name if inspect.isfunction(fn_name) else fn_name()
else:
raise ValueError('Unknown activation function/ class')
except Exception as e:
print(e)
return None
def get_loss(loss_name):
if loss_name is None:
return None
loss_modules = ['trident.optims.tensorflow_losses']
if loss_name in __all__:
loss_fn = get_class(loss_name, loss_modules)
else:
try:
loss_fn = get_class(camel2snake(loss_name), loss_modules)
except Exception:
loss_fn = get_class(loss_name, loss_modules)
return loss_fn
def get_normalization(fn_name):
if fn_name is None:
return None
elif isinstance(fn_name, Layer) and 'Norm' in fn_name.__class__.__name__:
return fn_name
elif inspect.isclass(fn_name) and fn_name.__class__.__name__ == type:
return fn_name()
elif inspect.isclass(fn_name):
return fn_name
elif isinstance(fn_name, str):
if fn_name.lower().strip() in ['instance_norm', 'instance', 'in', 'i']:
return InstanceNorm()
elif fn_name.lower().strip() in ['batch_norm', 'batch', 'bn', 'b']:
return BatchNorm()
elif fn_name.lower().strip() in ['layer_norm', 'layer', 'ln', 'l']:
return BatchNorm()
elif fn_name.lower().strip() in ['group_norm', 'group', 'gn', 'g']:
return GroupNorm(num_groups=16)
elif fn_name.lower().strip() in ['evo_normb0', 'evo-b0', 'evob0']:
return EvoNormB0()
elif fn_name.lower().strip() in ['evo_norms0', 'evo-s0', 'evos0']:
return EvoNormS0()
elif fn_name.lower().strip() in [
'spectral_norm', 'spectral', 'spec', 'sp', 's'
]:
return SpectralNorm
elif fn_name.lower().strip() in ['l2_norm', 'l2']:
return L2Norm()
fn_modules = ['trident.layers.pytorch_normalizations']
normalization_fn_ = get_class(fn_name, fn_modules)
normalization_fn = normalization_fn_
return normalization_fn
def get_normalization(fn_name):
if fn_name is None:
return None
elif isinstance(fn_name, Layer) and 'Norm' in fn_name.__class__.__name__:
return fn_name
elif inspect.isclass(fn_name):
return fn_name
elif isinstance(fn_name, str):
if fn_name.lower().strip() in ['instance', 'in', 'i']:
return None
#return InstanceNorm()
elif fn_name.lower().strip() in ['batch_norm', 'batch', 'bn', 'b']:
return BatchNorm2d()
elif fn_name.lower().strip() in ['group', 'g']:
return None
#return GroupNorm(num_groups=16)
elif inspect.isclass(fn_name):
return fn_name
fn_modules = ['trident.layers.tensorflow_normalizations']
normalization_fn_ = get_class(fn_name, fn_modules)
normalization_fn = normalization_fn_
return normalization_fn
