def plot_samples(epoch=None):
epoch = epoch or loop.epoch
with tk.layers.scoped_eval_mode(vae), T.no_grad():
logits = vae.p(n_z=100)['x'].distribution.logits
images = T.reshape(
T.cast(T.clip(T.nn.sigmoid(logits) * 255., 0., 255.), dtype=T.uint8),
[-1, 28, 28],
)
utils.save_images_collection(
images=T.to_numpy(images),
filename=exp.abspath(f'plotting/{epoch}.png'),
grid_size=(10, 10),
)
def test_get_activation_class(self):
x = T.random.randn([2, 3, 4])
for origin_name, factory, args, kwargs, expected in [
('Linear', None, None, None, None),
('ReLU', tk.layers.ReLU, (), {}, T.nn.relu(x)),
('Leaky_ReLU', tk.layers.LeakyReLU, (), {}, T.nn.leaky_relu(x)),
('Leaky_ReLU', tk.layers.LeakyReLU, (0.2, ), {},
T.nn.leaky_relu(x, 0.2)),
('Leaky_ReLU', tk.layers.LeakyReLU, (), {
'negative_slope': 0.2
}, T.nn.leaky_relu(x, 0.2)),
('Sigmoid', tk.layers.Sigmoid, (), {}, T.nn.sigmoid(x)),
('Tanh', tk.layers.Tanh, (), {}, T.tanh(x)),
('HardTanh', tk.layers.HardTanh, (), {}, T.clip(x, -1., 1.)),
('HardTanh', tk.layers.HardTanh, (-2., 3.), {}, T.clip(x, -2.,
3.)),
('HardTanh', tk.layers.HardTanh, (), {
'min_val': -2.,
'max_val': 3.
}, T.clip(x, -2., 3.)),
('Log_Softmax', tk.layers.LogSoftmax, (), {}, T.nn.log_softmax(x)),
]:
name_candidates = (None, ) if origin_name is None else (
origin_name, origin_name.lower(), origin_name.replace('_', ''),
origin_name.replace('_', '').lower())
for name in name_candidates:
err_msg = f'{name}, {factory}, {args}, {kwargs}, {expected}'
self.assertIs(tk.layers.get_activation_class(name), factory)
if factory is not None:
assert_allclose(factory(*args, **kwargs)(x),
expected,
err_msg=err_msg)
# unsupported activation
with pytest.raises(ValueError,
match='Unsupported activation: invalid'):
_ = tk.layers.get_activation_class('invalid')
def naive_clip_by_value(grads, clip_min, clip_max):
return [T.clip(g, clip_min, clip_max) for g in grads]