tensorflow - How to better organize the nodes in tensorboard with keras? -

i'm using keras instead of dealing tensorflow because simplicity. when tried visiualize computational graph in keras sending keras.callbacks.tensorboard instance model.fit()'s callbacks argument. graph got tensorboard awkward, demonstration purpose, here build simple linear classifier 1 unit in 1 dense layer. graph looks this:

could same thing did in tensorflow, use name_space group things , give layers, bias, weights names? mean, in graph here, it's such mess, can understand dense layer, , logistic loss namespace. typically tensorflow, can see train namespace, , not many nodes without namespace here. how can make more clear?

tensorflow graph shows computations being called. won't able simplify it.

as alternative, keras has it's own layer-by-layer graph. shows clear , concise structure of network. can generate calling

from keras.utils import plot_model plot_model(model, to_file='/some/pathname/model.png') 

last, can call model.summary(), generate textual version of graph, additional summaries.

here output of model.summary() example:

layer (type)                     output shape          param #     connected                      ==================================================================================================== input_1 (inputlayer)             (none, 2048)          0                                             ____________________________________________________________________________________________________ activation_1 (activation)        (none, 2048)          0                                             ____________________________________________________________________________________________________ dense_1 (dense)                  (none, 511)           1047039                                       ____________________________________________________________________________________________________ activation_2 (activation)        (none, 511)           0                                             ____________________________________________________________________________________________________ decoder_layer_1 (decoderlayer)   (none, 512)           0                                             ____________________________________________________________________________________________________ ctg_output (orlayer)             (none, 201)           102912                                        ____________________________________________________________________________________________________ att_output (orlayer)             (none, 312)           159744                                        ==================================================================================================== total params: 1,309,695.0 trainable params: 1,309,695.0 non-trainable params: 0.0