0

I am trying to implement DenseNet-121 model using tensorflow. I have already implemented the data pipeline and have already pre-processed the data which I have indicated by "some codes here" comment due to it being a bit lengthy(Will upload if needed) My code snippet is as follows:

import tensorflow as tf
 ''' Some Codes Here '''

from tensorflow.keras.layers import Input, Conv2D, BatchNormalization, Dense, AvgPool2D, GlobalAveragePooling2D, MaxPool2D, ReLU, concatenate

# Creating Densenet121
def densenet(input_shape, n_classes, filters = 32):
    
    #batch norm + relu + conv
    def bn_rl_conv(x,filters,kernel=1,strides=1):
        
        x = BatchNormalization()(x)
        x = ReLU()(x)
        x = Conv2D(filters, kernel, strides=strides,padding = 'same')(x)
        return x
    
    def dense_block(x, repetition):
        
        for _ in range(repetition):
            y = bn_rl_conv(x, 4*filters)
            y = bn_rl_conv(y, filters, 3)
            x = concatenate([y,x])
        return x
        
    def transition_layer(x):
        
        x = bn_rl_conv(x, tf.keras.backend.int_shape(x)[-1] //2 )
        x = AvgPool2D(2, strides = 2, padding = 'same')(x)
        return x
    
    inp = Input (input_shape)
    x = Conv2D(64, 7, padding = 'same', input_shape=input_shape)(inp)
    x = MaxPool2D(3, strides = 2, padding = 'same')(x)
    
    for repetition in [6,12,24,16]:
        
        d = dense_block(x, repetition)
        x = transition_layer(d)
    x = GlobalAveragePooling2D()(d)
    output = Dense(n_classes, activation = tf.nn.relu)(x)
    
    model = tf.keras.Model(inp, output)     # Line of error
    return model


input_shape = (224, 224, 3)
n_classes = 3
model1 = densenet(input_shape,n_classes)
model1.add(Dense(1, activation='sigmoid'))
model1.summary()

The error I am receiving from the above line marked by comment is as follows:

  File "C:\Users\USER\Ss\Image_Classification using DenseNet\main.py", line 131, in <module>
    model1 = densenet(input_shape,n_classes)

  File "C:\Users\USER\Ss\Image_Classification using DenseNet\main.py", line 125, in densenet
    model = tf.keras.Model(inp, output)     # Line of error

  File "C:\Users\USER\AppData\Local\Programs\Python\Python39\Lib\site-packages\keras\engine\training.py", line 191, in __new__
    from keras.engine import functional

  File "C:\Users\USER\AppData\Local\Programs\Python\Python39\Lib\site-packages\keras\__init__.py", line 20, in <module>
    from keras import distribute

  File "C:\Users\USER\AppData\Local\Programs\Python\Python39\Lib\site-packages\keras\distribute\__init__.py", line 18, in <module>
    from keras.distribute import sidecar_evaluator

  File "C:\Users\USER\AppData\Local\Programs\Python\Python39\Lib\site-packages\keras\distribute\sidecar_evaluator.py", line 22, in <module>
    from keras.optimizers.optimizer_experimental import (

  File "C:\Users\USER\AppData\Local\Programs\Python\Python39\Lib\site-packages\keras\optimizers\__init__.py", line 26, in <module>
    from keras.optimizers.legacy import adadelta as adadelta_legacy

  File "C:\Users\USER\AppData\Local\Programs\Python\Python39\Lib\site-packages\keras\optimizers\legacy\adadelta.py", line 17, in <module>
    from keras.optimizers.optimizer_v2 import adadelta

  File "C:\Users\USER\AppData\Local\Programs\Python\Python39\Lib\site-packages\keras\optimizers\optimizer_v2\adadelta.py", line 21, in <module>
    from keras.optimizers.optimizer_v2 import optimizer_v2

  File "C:\Users\USER\AppData\Local\Programs\Python\Python39\Lib\site-packages\keras\optimizers\optimizer_v2\optimizer_v2.py", line 38, in <module>
    keras_optimizers_gauge = tf.__internal__.monitoring.BoolGauge(

  File "C:\Users\USER\AppData\Local\Programs\Python\Python39\Lib\site-packages\tensorflow\python\eager\monitoring.py", line 356, in __init__
    super(BoolGauge, self).__init__('BoolGauge', _bool_gauge_methods,

  File "C:\Users\USER\AppData\Local\Programs\Python\Python39\Lib\site-packages\tensorflow\python\eager\monitoring.py", line 131, in __init__
    self._metric = self._metric_methods[self._label_length].create(*args)

AlreadyExistsError: Another metric with the same name already exists.


2022-10-27 01:01:14.175371: E tensorflow/core/lib/monitoring/collection_registry.cc:81] Cannot register 2 metrics with the same name: /tensorflow/api/keras/optimizers

Tensorflow version: 2.10.0

Keras version: 2.10.0

Please Help.

Edit: Found that after its execution and error in spyder, if I type import keras in terminal, it generates the same error, while if I type the same in say another IDE terminal alone, it generates no error during importing.

Surya Majumder
  • 95
  • 10
  • Uninstall keras. Tensorflow will naturally install Keras, and you can use tf.keras. – Vijay Mariappan Oct 27 '22 at 08:00
  • -> model1.add(Dense(1, activation='sigmoid')) Remove this line from your code... Once a model is built you cannot add a layer to it unless you can make another TensorFlow model and then add layer to it. – Mohammad Ahmed Oct 27 '22 at 13:48
  • Hi @SuryaMajumder, please try executing the same in the latest tf-nightly version also kindly refer to this similar [issue](https://stackoverflow.com/questions/69694944/how-to-fix-error-cannot-register-2-metrics-with-the-same-name-tensorflow-api) and let us know if the issue still persists. Thank you! –  Nov 07 '22 at 14:31

0 Answers0