TensorFlow is open-source python library designed by Google to develop Machine Learning models and deep learning neural networks. argmax() is a method present in tensorflow math module. This method is used to find the maximum value across the axes.
Syntax: tensorflow.math.argmax( input,axes,output_type,name ) Arguments: 1. input: It is a tensor. Allowed dtypes for this tensor are float32, float64, int32, uint8, int16, int8, complex64, int64, qint8, quint8, qint32, bfloat16, uint16, complex128, half, uint32, uint64. 2. axes: It is also a vector. It describes the axes to reduce the tensor. Allowed dtype are int32 and int64. Also [-rank(input),rank(input)) is the range allowed. axes=0 is used for vector. 3. output_type: It defines the dtype in which returned result should be. Allowed values are int32, int64 and the default value is int64. 4. name: It is an optional argument which defines name for the operation. Return: A tensor of output_type which contains the indices of the maximum value along the axes.
Example 1:
Python3
# importing the library import tensorflow as tf # initializing the constant tensor a = tf.constant([ 5 , 10 , 5.6 , 7.9 , 1 , 50 ]) # 50 is the maximum value at index 5 # getting the maximum value index tensor b = tf.math.argmax( input = a) # printing the tensor print ( 'tensor: ' ,b) # Evaluating the value of tensor c = tf.keras.backend. eval (b) #printing the value print ( 'value: ' ,c) |
Output:
tensor: tf.Tensor(5, shape=(), dtype=int64) value: 5
Example 2:
This example uses a tensor of shape(3,3).
Python3
# importing the library import tensorflow as tf # initializing the constant tensor a = tf.constant(value = [ 9 , 8 , 7 , 3 , 5 , 4 , 6 , 2 , 1 ],shape = ( 3 , 3 )) # printing the initialized tensor print (a) # getting the maximum value indices tensor b = tf.math.argmax( input = a) # printing the tensor print ( 'Indices Tensor: ' ,b) # Evaluating the tensor value c = tf.keras.backend. eval (b) # printing the value print ( 'Indices: ' ,c) |
Output:
tf.Tensor( [[9 8 7] [3 5 4] [6 2 1]], shape=(3, 3), dtype=int32) Indices tensor: tf.Tensor([0 0 0], shape=(3,), dtype=int64) Indices: [0 0 0] # maximum value along the axes are 9,8,7 at indices 0,0,0 respectively.