SHARE
TWEET

Python - keras

furas Jun 9th, 2018 (edited) 75 Never
Not a member of Pastebin yet? Sign Up, it unlocks many cool features!
  1. import keras as ks
  2. import numpy as np
  3.  
  4. features = np.zeros((1,13,22))
  5. print('features:', features.shape)
  6.  
  7. input_size = features.shape[1:]   #input_size (13, 22)
  8. print('input_size:', input_size)
  9.  
  10. # input layer
  11. input_layer = ks.layers.Input(shape=(*input_size,), name='input')
  12. print('input:', input_layer.shape) # (?, 13, 22)
  13.  
  14. # noise
  15. x = ks.layers.GaussianNoise(stddev=0.1)(input_layer)  
  16. print('1', x.shape) # (?, 13, 22)
  17.  
  18. # conv layer
  19. x = ks.layers.Conv1D(filters=8, kernel_size=3, strides=1, activation=ks.activations.relu, padding='same')(x)
  20. print('2', x.shape) # (?, 13, 8)
  21.  
  22. x = ks.layers.ZeroPadding1D(padding=(1, 0))(x)
  23. print('3', x.shape) # (?, 14, 8)
  24.  
  25. x = ks.layers.MaxPool1D(pool_size=2, strides=None)(x)
  26. print('4', x.shape) # (?, 7, 8)
  27.  
  28. x = ks.layers.Conv1D(filters=8, kernel_size=3, strides=1, activation=ks.activations.relu, padding='same')(x)
  29. print('5', x.shape) # (?, 7, 8)
  30.  
  31. x = ks.layers.UpSampling1D(size=2)(x)
  32. print('6', x.shape) # (?, 14, 8)
  33.  
  34. x = ks.layers.Conv1D(filters=8, kernel_size=3, strides=1, activation=ks.activations.relu, padding='same')(x)
  35. print('7', x.shape) # (?, 14, 8)
  36.  
  37. x = ks.layers.Conv1D(filters=input_size[-1], kernel_size=3, strides=1, activation=ks.activations.relu, padding='same')(x)
  38. print('8', x.shape) # (?, 14, 22)
  39.  
  40. # output
  41. output_layer = x
  42. print('output', output_layer.shape)
  43.  
  44. model = ks.models.Model(inputs=input_layer, outputs=output_layer)
  45. model.compile(optimizer='rmsprop',
  46.               loss='categorical_crossentropy',
  47.               metrics=['accuracy'])
  48.  
  49. print(model.summary())
  50.  
  51. labels = np.zeros((1,13,22)) # gives error
  52. #labels = np.zeros((1,14,22)) # works correctly
  53. model.fit(features, labels)  # <--- # ValueError: Error when checking target: expected conv1d_4 to have shape (14, 22) but got array with shape (13, 22)
RAW Paste Data
We use cookies for various purposes including analytics. By continuing to use Pastebin, you agree to our use of cookies as described in the Cookies Policy. OK, I Understand
 
Top