Functions — github.com/conan7882/adversarial-autoencoders

CodeHub MCPcopy Create free account

hub / github.com/conan7882/adversarial-autoencoders / functions

Functions121 in github.com/conan7882/adversarial-autoencoders

⨍ Functions 121◇ Types & classes 7

Methodget_train_op

src/models/base.py:36

Methodget_train_summary

src/models/aae.py:272

Methodget_valid_summary

src/models/aae.py:260

Functionidentity

src/dataflow/mnist.py:15

Functioninterpolate_gm

(plot_size=20, interpolate_range=[-1., 1., -0.2, 0.2], mode_id=0, n_mode=10)

src/models/distribution.py:23

Functionleaky_relu

leaky_relu Allow a small non-zero gradient when the unit is not active Args: x (tf.tensor): a tensor leak (float):

src/models/layers.py:59

(out_dim, layer_dict=None, inputs=None, init_w=None, init_b=tf.zer

src/models/layers.py:73

Functionmake_list

src/utils/utils.py:7

Functionmax_pool

Max pooling layer Args: x (tf.tensor): a tensor name (str): name scope of the layer filter_size (int or list with

src/models/layers.py:171

Functionpreprocess_im

normalize input image to [-1., 1.]

experiment/aae_mnist.py:74

Functionpreprocess_im

experiment/vae_mnist.py:53

Methodreset_epoch

src/dataflow/mnist.py:148

Functionsoftmax

(logits, axis=-1, name='softmax')

src/models/layers.py:51

Functionsoftplus

src/models/layers.py:48

Functiontf_sample_diag_guassian

(mean, std, b_size, n_code)

src/models/ops.py:19

Functiontf_sample_standard_diag_guassian

(b_size, n_code)

src/models/ops.py:10

Functiontrain_by_cross_entropy_loss

(logits, labels, loss_weight, opt, var_list, name)

src/models/modules.py:136

Functiontrain_discrimator

(fake_in, real_in, loss_weight, opt, var_list, name)

src/models/modules.py:95

Functiontrain_generator

(fake_in, loss_weight, opt, var_list, name)

src/models/modules.py:121

Functiontranspose_conv

( filter_size, out_dim, layer_dict,

src/models/layers.py:112

Functionviz_batch_im

(batch_im, grid_size, save_path, gap=0, gap_color=0, shuffle=False)

src/utils/viz.py:10

← previous101–121 of 121, ranked by callers