Keras How to get the output of each layer

Question

I have trained a binary classification model with CNN  and here is my code  model   Sequential   model add Convolution2D nb filters  kernel size 0   kernel size 1                           border mode  valid                           input shape input shape   model add Activation  relu    model add Convolution2D nb filters  kernel size 0   kernel size 1    model add Activation  relu    model add MaxPooling2D pool size pool size      16  16  32  model add Convolution2D nb filters 2  kernel size 0   kernel size 1    model add Activation  relu    model add Convolution2D nb filters 2  kernel size 0   kernel size 1    model add Activation  relu    model add MaxPooling2D pool size pool size      8  8  64     2048  model add Flatten    model add Dense 1024   model add Activation  relu    model add Dropout 0 5   model add Dense 2      define a binary classification problem model add Activation  softmax     model compile loss  categorical crossentropy                 optimizer  adadelta                 metrics   accuracy    model fit x train  y train            batch size batch size            nb epoch nb epoch            verbose 1            validation data  x test  y test     And here  I wanna get the output of each layer just like TensorFlow  how can I do that

User · Answer

In case you have one of the following cases:

error: InvalidArgumentError: input_X:Y is both fed and fetched
case of multiple inputs

You need to do the following changes:

add filter out for input layers in outputs variable
minnor change on functors loop

Minimum example:

from keras.engine.input_layer import InputLayer
inp = model.input
outputs = [layer.output for layer in model.layers if not isinstance(layer, InputLayer)]
functors = [K.function(inp + [K.learning_phase()], [x]) for x in outputs]
layer_outputs = [fun([x1, x2, xn, 1]) for fun in functors]

User · Answer

From https   keras io getting-started faq  how-can-i-obtain-the-output-of-an-intermediate-layer  One simple way is to create a new Model that will output the layers that you are interested in   from keras models import Model  model          include here your original model  layer name    my layer  intermediate layer model   Model inputs model input                                   outputs model get layer layer name  output  intermediate output   intermediate layer model predict data    Alternatively  you can build a Keras function that will return the output of a certain layer given a certain input  for example   from keras import backend as K    with a Sequential model get 3rd layer output   K function  model layers 0  input                                      model layers 3  output   layer output   get 3rd layer output  x   0

User · Answer

Based on all the good answers of this thread  I wrote a library to fetch the output of each layer  It abstracts all the complexity and has been designed to be as user-friendly as possible   https   github com philipperemy keract  It handles almost all the edge cases  Hope it helps

User · Answer

Wanted to add this as a comment  but don t have high enough rep   to  indraforyou s answer to correct for the issue mentioned in  mathtick s comment  To avoid the InvalidArgumentError  input X Y is both fed and fetched  exception  simply replace the line outputs    layer output for layer in model layers  with outputs    layer output for layer in model layers  1    i e   adapting indraforyou s minimal working example   from keras import backend as K  inp   model input                                             input placeholder outputs    layer output for layer in model layers  1            all layer outputs except first  input  layer functor   K function  inp  K learning phase     outputs       evaluation function    Testing test   np random random input shape  np newaxis      layer outs   functor  test  1    print layer outs   p s  my attempts trying things such as outputs    layer output for layer in model layers 1    did not work

User · Answer

I wrote this function for myself  in Jupyter  and it was inspired by indraforyou s answer  It will plot all the layer outputs automatically  Your images must have a  x  y  1  shape where 1 stands for 1 channel  You just call plot layer outputs      to plot    matplotlib inline import matplotlib pyplot as plt from keras import backend as K  def get layer outputs        test image   YOUR IMAGE GOES HERE        outputs       layer output for layer in model layers             all layer outputs     comp graph    K function  model input    K learning phase      output   for output in outputs     evaluation functions        Testing     layer outputs list    op  test image  1    for op in comp graph      layer outputs           for layer output in layer outputs list          print layer output 0  0  shape  end   n------------------- n           layer outputs append layer output 0  0        return layer outputs  def plot layer outputs layer number           layer outputs   get layer outputs        x max   layer outputs layer number  shape 0      y max   layer outputs layer number  shape 1      n       layer outputs layer number  shape 2       L          for i in range n           L append np zeros  x max  y max         for i in range n           for x in range x max               for y in range y max                   L i  x  y    layer outputs layer number  x  y  i        for img in L          plt figure           plt imshow img  interpolation  nearest

User · Answer

Assuming you have   1- Keras pre-trained model    2- Input x as image or set of images  The resolution of image should be compatible with dimension of the input layer  For example 80 80 3 for 3-channels  RGB  image    3- The name of the output layer to get the activation  For example   flatten 2  layer  This should be include in the layer names variable  represents name of layers of the given model   4- batch size is an optional argument   Then you can easily use get activation function to get the activation of the output layer for a given input x and pre-trained model   import six import numpy as np import keras backend as k from numpy import float32 def get activations x  model  layer  batch size 128       Return the output of the specified layer for input  x    layer  is specified by layer index  between 0 and  nb layers - 1   or by name  The number of layers can be determined by counting the results returned by calling  layer names    param x  Input for computing the activations   type x   np ndarray   Example  x shape    80  80  3   param model  pre-trained Keras model  Including weights   type model  keras engine sequential Sequential  Example  model input shape    None  80  80  3   param layer  Layer for computing the activations  type layer   int  or  str   Example  layer    flatten 2   param batch size  Size of batches   type batch size   int   return  The output of  layer   where the first dimension is the batch size corresponding to  x    rtype   np ndarray   Example  activations shape    1  2000           layer names    layer name for layer in model layers      if isinstance layer  six string types           if layer not in layer names              raise ValueError  Layer name  s is not part of the graph     layer          layer name   layer     elif isinstance layer  int           if layer  lt  0 or layer  gt   len layer names               raise ValueError  Layer index  d is outside of range  0 to  d included                                    layer  len layer names  - 1           layer name   layer names layer      else          raise TypeError  Layer must be of type  str  or  int          layer output   model get layer layer name  output     layer input   model input     output func   k function  layer input    layer output          Apply preprocessing     if x shape    k int shape model input  1            x preproc   np expand dims x  0      else          x preproc   x     assert len x preproc shape     4        Determine shape of expected output and prepare array     output shape   output func  x preproc 0  None         0  shape     activations   np zeros  x preproc shape 0      output shape 1    dtype float32         Get activations with batching     for batch index in range int np ceil x preproc shape 0    float batch size              begin  end   batch index   batch size  min  batch index   1    batch size  x preproc shape 0           activations begin end    output func  x preproc begin end    0       return activations

User · Answer

Previous solutions were not working for me  I handled this issue as shown below  layer outputs      for i in range 1  len model layers        tmp model   Model model layers 0  input  model layers i  output      tmp output   tmp model predict img  0      layer outputs append tmp output

User · Answer

From  https   github com philipperemy keras-visualize-activations blob master read activations py  import keras backend as K  def get activations model  model inputs  print shape only False  layer name None       print  ----- activations -----       activations          inp   model input      model multi inputs cond   True     if not isinstance inp  list             only one input  let s wrap it in a list          inp    inp          model multi inputs cond   False      outputs    layer output for layer in model layers if                layer name    layer name or layer name is None     all layer outputs      funcs    K function inp    K learning phase      out   for out in outputs     evaluation functions      if model multi inputs cond          list inputs              list inputs extend model inputs          list inputs append 0       else          list inputs    model inputs  0          Learning phase  0   Test mode  no dropout or batch normalization        layer outputs    func  model inputs  0    0  for func in funcs      layer outputs    func list inputs  0  for func in funcs      for layer activations in layer outputs          activations append layer activations          if print shape only              print layer activations shape          else              print layer activations      return activations

User · Answer

You can easily get the outputs of any layer by using  model layers index  output  For all layers use this   from keras import backend as K  inp   model input                                             input placeholder outputs    layer output for layer in model layers             all layer outputs functors    K function  inp  K learning phase      out   for out in outputs       evaluation functions    Testing test   np random random input shape  np newaxis      layer outs    func  test  1    for func in functors  print layer outs   Note  To simulate Dropout use learning phase as 1  in layer outs otherwise use 0   Edit   based on comments   K function creates theano tensorflow tensor functions which is later used to get the output from the symbolic graph given the input    Now K learning phase   is required as an input as many Keras layers like Dropout Batchnomalization depend on it to change behavior during training and test time    So if you remove the dropout layer in your code you can simply use   from keras import backend as K  inp   model input                                             input placeholder outputs    layer output for layer in model layers             all layer outputs functors    K function  inp    out   for out in outputs       evaluation functions    Testing test   np random random input shape  np newaxis      layer outs    func  test   for func in functors  print layer outs   Edit 2  More optimized  I just realized that the previous answer is not that optimized as for each function evaluation the data will be transferred CPU- GPU memory and also the tensor calculations needs to be done for the lower layers over-n-over    Instead this is a much better way as you don t need multiple functions but a single function giving you the list of all outputs   from keras import backend as K  inp   model input                                             input placeholder outputs    layer output for layer in model layers             all layer outputs functor   K function  inp  K learning phase     outputs       evaluation function    Testing test   np random random input shape  np newaxis      layer outs   functor  test  1    print layer outs

User · Answer

This answer is based on  https   stackoverflow com a 59557567 2585501 To print the output of a single layer  from tensorflow keras import backend as K layerIndex   1 func   K function  model get layer index 0  input   model get layer index layerIndex  output  layerOutput   func  input data      input data is a numpy array print layerOutput   To print output of every layer  from tensorflow keras import backend as K for layerIndex  layer in enumerate model layers       func   K function  model get layer index 0  input   layer output      layerOutput   func  input data      input data is a numpy array     print layerOutput

User · Answer

Well  other answers are very complete  but there is a very basic way to  see   not to  get  the shapes   Just do a model summary    It will print all layers and their output shapes   None  values will indicate variable dimensions  and the first dimension will be the batch size

User · Answer

Following looks very simple to me   model layers idx  output   Above is a tensor object  so you can modify it using operations that can be applied to a tensor object    For example  to get the shape model layers idx  output get shape    idx is the index of the layer and you can find it from model summary

[python] Keras, How to get the output of each layer?

Examples related to python

Examples related to tensorflow

Examples related to deep-learning

Examples related to keras