Why is the Keras model always predicting the same class / How can I improve the accuracy of this model?












0












$begingroup$


First post here. I'm working on a project about multi-class image classification and created a python script using Keras to train a model with transfer learning. To my dismay the model has always predicted the same class, I've simplified the model down to 3 image classes (I'm using a kaggle food image stock with 800 training samples and 800 validation samples per class plus image reformatting) and tried different optimizers, yet it still comes down to the same class while the model also apparently only has an accuracy of ~0.2563 at 25 epochs of training. I've posted the code below, how can I improve the accuracy of this script and solve the same predicted class problem?



import pandas as pd
import numpy as np
import os
import keras
import matplotlib.pyplot as plt
from keras.layers import Dense, GlobalAveragePooling2D
from keras.preprocessing import image
from keras.preprocessing.image import ImageDataGenerator
from keras.models import Model
from keras import optimizers
from keras import applications
from keras.applications.vgg16 import preprocess_input

img_classes = 3

base_model = applications.VGG16(weights='imagenet', include_top=False)

x = base_model.output
x = GlobalAveragePooling2D()(x)
x = Dense(1024, activation='relu')(x)
x = Dense(1024, activation='relu')(x)
x = Dense(512, activation='relu')(x)
preds = Dense(img_classes, activation='softmax')(x)

model = Model(inputs=base_model.input, outputs=preds)

for i, layer in enumerate(model.layers):
print(i, layer.name)

for layer in model.layers[:25]:
layer.trainable = False

train_datagen = ImageDataGenerator(rescale=1./255,
rotation_range=40,
width_shift_range=0.2,
height_shift_range=0.2,
shear_range=0.2,
zoom_range=0.2,
horizontal_flip=True,
fill_mode='nearest',
preprocessing_function=preprocess_input)

train_generator = train_datagen.flow_from_directory('./food-101/bigtrain',
target_size=(128, 128),
color_mode='rgb',
classes=['apple_pie', 'churros', 'miso_soup'],
batch_size=1,
class_mode='categorical',
shuffle=True)

val_datagen = ImageDataGenerator(
rescale=1./255,
rotation_range=40,
width_shift_range=0.2,
height_shift_range=0.2,
shear_range=0.2,
zoom_range=0.2,
horizontal_flip=True,
fill_mode='nearest',
preprocessing_function=preprocess_input,)

val_generator = val_datagen.flow_from_directory(
'./food-101/bigval',
target_size=(128, 128),
classes=['apple_pie', 'churros', 'miso_soup'],
batch_size=1,
class_mode='categorical',
shuffle=True)

# model.compile(optimizer='rmsprop', loss='categorical_crossentropy', metrics=['accuracy'])

model.compile(optimizer=optimizers.SGD(lr=0.00001,
momentum=0.9,
decay=0.0001,
nesterov=True), loss='categorical_crossentropy', metrics=['accuracy'])

batch_size = 1


validation_steps = 64 // batch_size
step_size_train = train_generator.n//train_generator.batch_size

model.fit_generator(generator=train_generator,
steps_per_epoch=step_size_train,
epochs=25,
validation_data=val_generator,
validation_steps=validation_steps)

model.save('./test_try_vgg_9.h5')


Example prediction results:



classes: apple_pie, churros, miso_soup



miso soup
[0.3202575 0.48074356 0.19899891] rmsprop
[0.45246536 0.4505403 0.09699439] sgd

churros
[0.37473327 0.35784692 0.2674198 ] rmsprop
[0.4145825 0.465228 0.12018944] sgd


This is the prediction script:



from keras.models import load_model
from keras import optimizers
from keras.preprocessing import image
import numpy as np
from keras.applications.vgg16 import preprocess_input

# dimensions of our images
img_width, img_height = 512, 512

# load model
model = load_model('./test_try_vgg_9.h5')

# predicting images
img = image.load_img('./food-101/training/apple_pie/551535.jpg')
x = image.img_to_array(img)
x = np.expand_dims(x, axis=0)
x = preprocess_input(x)

pred = model.predict(x)
print("Probability: ")
print(pred[0])









share|improve this question







New contributor




vanillinx is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.







$endgroup$

















    0












    $begingroup$


    First post here. I'm working on a project about multi-class image classification and created a python script using Keras to train a model with transfer learning. To my dismay the model has always predicted the same class, I've simplified the model down to 3 image classes (I'm using a kaggle food image stock with 800 training samples and 800 validation samples per class plus image reformatting) and tried different optimizers, yet it still comes down to the same class while the model also apparently only has an accuracy of ~0.2563 at 25 epochs of training. I've posted the code below, how can I improve the accuracy of this script and solve the same predicted class problem?



    import pandas as pd
    import numpy as np
    import os
    import keras
    import matplotlib.pyplot as plt
    from keras.layers import Dense, GlobalAveragePooling2D
    from keras.preprocessing import image
    from keras.preprocessing.image import ImageDataGenerator
    from keras.models import Model
    from keras import optimizers
    from keras import applications
    from keras.applications.vgg16 import preprocess_input

    img_classes = 3

    base_model = applications.VGG16(weights='imagenet', include_top=False)

    x = base_model.output
    x = GlobalAveragePooling2D()(x)
    x = Dense(1024, activation='relu')(x)
    x = Dense(1024, activation='relu')(x)
    x = Dense(512, activation='relu')(x)
    preds = Dense(img_classes, activation='softmax')(x)

    model = Model(inputs=base_model.input, outputs=preds)

    for i, layer in enumerate(model.layers):
    print(i, layer.name)

    for layer in model.layers[:25]:
    layer.trainable = False

    train_datagen = ImageDataGenerator(rescale=1./255,
    rotation_range=40,
    width_shift_range=0.2,
    height_shift_range=0.2,
    shear_range=0.2,
    zoom_range=0.2,
    horizontal_flip=True,
    fill_mode='nearest',
    preprocessing_function=preprocess_input)

    train_generator = train_datagen.flow_from_directory('./food-101/bigtrain',
    target_size=(128, 128),
    color_mode='rgb',
    classes=['apple_pie', 'churros', 'miso_soup'],
    batch_size=1,
    class_mode='categorical',
    shuffle=True)

    val_datagen = ImageDataGenerator(
    rescale=1./255,
    rotation_range=40,
    width_shift_range=0.2,
    height_shift_range=0.2,
    shear_range=0.2,
    zoom_range=0.2,
    horizontal_flip=True,
    fill_mode='nearest',
    preprocessing_function=preprocess_input,)

    val_generator = val_datagen.flow_from_directory(
    './food-101/bigval',
    target_size=(128, 128),
    classes=['apple_pie', 'churros', 'miso_soup'],
    batch_size=1,
    class_mode='categorical',
    shuffle=True)

    # model.compile(optimizer='rmsprop', loss='categorical_crossentropy', metrics=['accuracy'])

    model.compile(optimizer=optimizers.SGD(lr=0.00001,
    momentum=0.9,
    decay=0.0001,
    nesterov=True), loss='categorical_crossentropy', metrics=['accuracy'])

    batch_size = 1


    validation_steps = 64 // batch_size
    step_size_train = train_generator.n//train_generator.batch_size

    model.fit_generator(generator=train_generator,
    steps_per_epoch=step_size_train,
    epochs=25,
    validation_data=val_generator,
    validation_steps=validation_steps)

    model.save('./test_try_vgg_9.h5')


    Example prediction results:



    classes: apple_pie, churros, miso_soup



    miso soup
    [0.3202575 0.48074356 0.19899891] rmsprop
    [0.45246536 0.4505403 0.09699439] sgd

    churros
    [0.37473327 0.35784692 0.2674198 ] rmsprop
    [0.4145825 0.465228 0.12018944] sgd


    This is the prediction script:



    from keras.models import load_model
    from keras import optimizers
    from keras.preprocessing import image
    import numpy as np
    from keras.applications.vgg16 import preprocess_input

    # dimensions of our images
    img_width, img_height = 512, 512

    # load model
    model = load_model('./test_try_vgg_9.h5')

    # predicting images
    img = image.load_img('./food-101/training/apple_pie/551535.jpg')
    x = image.img_to_array(img)
    x = np.expand_dims(x, axis=0)
    x = preprocess_input(x)

    pred = model.predict(x)
    print("Probability: ")
    print(pred[0])









    share|improve this question







    New contributor




    vanillinx is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
    Check out our Code of Conduct.







    $endgroup$















      0












      0








      0





      $begingroup$


      First post here. I'm working on a project about multi-class image classification and created a python script using Keras to train a model with transfer learning. To my dismay the model has always predicted the same class, I've simplified the model down to 3 image classes (I'm using a kaggle food image stock with 800 training samples and 800 validation samples per class plus image reformatting) and tried different optimizers, yet it still comes down to the same class while the model also apparently only has an accuracy of ~0.2563 at 25 epochs of training. I've posted the code below, how can I improve the accuracy of this script and solve the same predicted class problem?



      import pandas as pd
      import numpy as np
      import os
      import keras
      import matplotlib.pyplot as plt
      from keras.layers import Dense, GlobalAveragePooling2D
      from keras.preprocessing import image
      from keras.preprocessing.image import ImageDataGenerator
      from keras.models import Model
      from keras import optimizers
      from keras import applications
      from keras.applications.vgg16 import preprocess_input

      img_classes = 3

      base_model = applications.VGG16(weights='imagenet', include_top=False)

      x = base_model.output
      x = GlobalAveragePooling2D()(x)
      x = Dense(1024, activation='relu')(x)
      x = Dense(1024, activation='relu')(x)
      x = Dense(512, activation='relu')(x)
      preds = Dense(img_classes, activation='softmax')(x)

      model = Model(inputs=base_model.input, outputs=preds)

      for i, layer in enumerate(model.layers):
      print(i, layer.name)

      for layer in model.layers[:25]:
      layer.trainable = False

      train_datagen = ImageDataGenerator(rescale=1./255,
      rotation_range=40,
      width_shift_range=0.2,
      height_shift_range=0.2,
      shear_range=0.2,
      zoom_range=0.2,
      horizontal_flip=True,
      fill_mode='nearest',
      preprocessing_function=preprocess_input)

      train_generator = train_datagen.flow_from_directory('./food-101/bigtrain',
      target_size=(128, 128),
      color_mode='rgb',
      classes=['apple_pie', 'churros', 'miso_soup'],
      batch_size=1,
      class_mode='categorical',
      shuffle=True)

      val_datagen = ImageDataGenerator(
      rescale=1./255,
      rotation_range=40,
      width_shift_range=0.2,
      height_shift_range=0.2,
      shear_range=0.2,
      zoom_range=0.2,
      horizontal_flip=True,
      fill_mode='nearest',
      preprocessing_function=preprocess_input,)

      val_generator = val_datagen.flow_from_directory(
      './food-101/bigval',
      target_size=(128, 128),
      classes=['apple_pie', 'churros', 'miso_soup'],
      batch_size=1,
      class_mode='categorical',
      shuffle=True)

      # model.compile(optimizer='rmsprop', loss='categorical_crossentropy', metrics=['accuracy'])

      model.compile(optimizer=optimizers.SGD(lr=0.00001,
      momentum=0.9,
      decay=0.0001,
      nesterov=True), loss='categorical_crossentropy', metrics=['accuracy'])

      batch_size = 1


      validation_steps = 64 // batch_size
      step_size_train = train_generator.n//train_generator.batch_size

      model.fit_generator(generator=train_generator,
      steps_per_epoch=step_size_train,
      epochs=25,
      validation_data=val_generator,
      validation_steps=validation_steps)

      model.save('./test_try_vgg_9.h5')


      Example prediction results:



      classes: apple_pie, churros, miso_soup



      miso soup
      [0.3202575 0.48074356 0.19899891] rmsprop
      [0.45246536 0.4505403 0.09699439] sgd

      churros
      [0.37473327 0.35784692 0.2674198 ] rmsprop
      [0.4145825 0.465228 0.12018944] sgd


      This is the prediction script:



      from keras.models import load_model
      from keras import optimizers
      from keras.preprocessing import image
      import numpy as np
      from keras.applications.vgg16 import preprocess_input

      # dimensions of our images
      img_width, img_height = 512, 512

      # load model
      model = load_model('./test_try_vgg_9.h5')

      # predicting images
      img = image.load_img('./food-101/training/apple_pie/551535.jpg')
      x = image.img_to_array(img)
      x = np.expand_dims(x, axis=0)
      x = preprocess_input(x)

      pred = model.predict(x)
      print("Probability: ")
      print(pred[0])









      share|improve this question







      New contributor




      vanillinx is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
      Check out our Code of Conduct.







      $endgroup$




      First post here. I'm working on a project about multi-class image classification and created a python script using Keras to train a model with transfer learning. To my dismay the model has always predicted the same class, I've simplified the model down to 3 image classes (I'm using a kaggle food image stock with 800 training samples and 800 validation samples per class plus image reformatting) and tried different optimizers, yet it still comes down to the same class while the model also apparently only has an accuracy of ~0.2563 at 25 epochs of training. I've posted the code below, how can I improve the accuracy of this script and solve the same predicted class problem?



      import pandas as pd
      import numpy as np
      import os
      import keras
      import matplotlib.pyplot as plt
      from keras.layers import Dense, GlobalAveragePooling2D
      from keras.preprocessing import image
      from keras.preprocessing.image import ImageDataGenerator
      from keras.models import Model
      from keras import optimizers
      from keras import applications
      from keras.applications.vgg16 import preprocess_input

      img_classes = 3

      base_model = applications.VGG16(weights='imagenet', include_top=False)

      x = base_model.output
      x = GlobalAveragePooling2D()(x)
      x = Dense(1024, activation='relu')(x)
      x = Dense(1024, activation='relu')(x)
      x = Dense(512, activation='relu')(x)
      preds = Dense(img_classes, activation='softmax')(x)

      model = Model(inputs=base_model.input, outputs=preds)

      for i, layer in enumerate(model.layers):
      print(i, layer.name)

      for layer in model.layers[:25]:
      layer.trainable = False

      train_datagen = ImageDataGenerator(rescale=1./255,
      rotation_range=40,
      width_shift_range=0.2,
      height_shift_range=0.2,
      shear_range=0.2,
      zoom_range=0.2,
      horizontal_flip=True,
      fill_mode='nearest',
      preprocessing_function=preprocess_input)

      train_generator = train_datagen.flow_from_directory('./food-101/bigtrain',
      target_size=(128, 128),
      color_mode='rgb',
      classes=['apple_pie', 'churros', 'miso_soup'],
      batch_size=1,
      class_mode='categorical',
      shuffle=True)

      val_datagen = ImageDataGenerator(
      rescale=1./255,
      rotation_range=40,
      width_shift_range=0.2,
      height_shift_range=0.2,
      shear_range=0.2,
      zoom_range=0.2,
      horizontal_flip=True,
      fill_mode='nearest',
      preprocessing_function=preprocess_input,)

      val_generator = val_datagen.flow_from_directory(
      './food-101/bigval',
      target_size=(128, 128),
      classes=['apple_pie', 'churros', 'miso_soup'],
      batch_size=1,
      class_mode='categorical',
      shuffle=True)

      # model.compile(optimizer='rmsprop', loss='categorical_crossentropy', metrics=['accuracy'])

      model.compile(optimizer=optimizers.SGD(lr=0.00001,
      momentum=0.9,
      decay=0.0001,
      nesterov=True), loss='categorical_crossentropy', metrics=['accuracy'])

      batch_size = 1


      validation_steps = 64 // batch_size
      step_size_train = train_generator.n//train_generator.batch_size

      model.fit_generator(generator=train_generator,
      steps_per_epoch=step_size_train,
      epochs=25,
      validation_data=val_generator,
      validation_steps=validation_steps)

      model.save('./test_try_vgg_9.h5')


      Example prediction results:



      classes: apple_pie, churros, miso_soup



      miso soup
      [0.3202575 0.48074356 0.19899891] rmsprop
      [0.45246536 0.4505403 0.09699439] sgd

      churros
      [0.37473327 0.35784692 0.2674198 ] rmsprop
      [0.4145825 0.465228 0.12018944] sgd


      This is the prediction script:



      from keras.models import load_model
      from keras import optimizers
      from keras.preprocessing import image
      import numpy as np
      from keras.applications.vgg16 import preprocess_input

      # dimensions of our images
      img_width, img_height = 512, 512

      # load model
      model = load_model('./test_try_vgg_9.h5')

      # predicting images
      img = image.load_img('./food-101/training/apple_pie/551535.jpg')
      x = image.img_to_array(img)
      x = np.expand_dims(x, axis=0)
      x = preprocess_input(x)

      pred = model.predict(x)
      print("Probability: ")
      print(pred[0])






      keras multiclass-classification






      share|improve this question







      New contributor




      vanillinx is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
      Check out our Code of Conduct.











      share|improve this question







      New contributor




      vanillinx is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
      Check out our Code of Conduct.









      share|improve this question




      share|improve this question






      New contributor




      vanillinx is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
      Check out our Code of Conduct.









      asked 14 mins ago









      vanillinxvanillinx

      1




      1




      New contributor




      vanillinx is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
      Check out our Code of Conduct.





      New contributor





      vanillinx is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
      Check out our Code of Conduct.






      vanillinx is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
      Check out our Code of Conduct.






















          0






          active

          oldest

          votes











          Your Answer





          StackExchange.ifUsing("editor", function () {
          return StackExchange.using("mathjaxEditing", function () {
          StackExchange.MarkdownEditor.creationCallbacks.add(function (editor, postfix) {
          StackExchange.mathjaxEditing.prepareWmdForMathJax(editor, postfix, [["$", "$"], ["\\(","\\)"]]);
          });
          });
          }, "mathjax-editing");

          StackExchange.ready(function() {
          var channelOptions = {
          tags: "".split(" "),
          id: "557"
          };
          initTagRenderer("".split(" "), "".split(" "), channelOptions);

          StackExchange.using("externalEditor", function() {
          // Have to fire editor after snippets, if snippets enabled
          if (StackExchange.settings.snippets.snippetsEnabled) {
          StackExchange.using("snippets", function() {
          createEditor();
          });
          }
          else {
          createEditor();
          }
          });

          function createEditor() {
          StackExchange.prepareEditor({
          heartbeatType: 'answer',
          autoActivateHeartbeat: false,
          convertImagesToLinks: false,
          noModals: true,
          showLowRepImageUploadWarning: true,
          reputationToPostImages: null,
          bindNavPrevention: true,
          postfix: "",
          imageUploader: {
          brandingHtml: "Powered by u003ca class="icon-imgur-white" href="https://imgur.com/"u003eu003c/au003e",
          contentPolicyHtml: "User contributions licensed under u003ca href="https://creativecommons.org/licenses/by-sa/3.0/"u003ecc by-sa 3.0 with attribution requiredu003c/au003e u003ca href="https://stackoverflow.com/legal/content-policy"u003e(content policy)u003c/au003e",
          allowUrls: true
          },
          onDemand: true,
          discardSelector: ".discard-answer"
          ,immediatelyShowMarkdownHelp:true
          });


          }
          });






          vanillinx is a new contributor. Be nice, and check out our Code of Conduct.










          draft saved

          draft discarded


















          StackExchange.ready(
          function () {
          StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fdatascience.stackexchange.com%2fquestions%2f44595%2fwhy-is-the-keras-model-always-predicting-the-same-class-how-can-i-improve-the%23new-answer', 'question_page');
          }
          );

          Post as a guest















          Required, but never shown

























          0






          active

          oldest

          votes








          0






          active

          oldest

          votes









          active

          oldest

          votes






          active

          oldest

          votes








          vanillinx is a new contributor. Be nice, and check out our Code of Conduct.










          draft saved

          draft discarded


















          vanillinx is a new contributor. Be nice, and check out our Code of Conduct.













          vanillinx is a new contributor. Be nice, and check out our Code of Conduct.












          vanillinx is a new contributor. Be nice, and check out our Code of Conduct.
















          Thanks for contributing an answer to Data Science Stack Exchange!


          • Please be sure to answer the question. Provide details and share your research!

          But avoid



          • Asking for help, clarification, or responding to other answers.

          • Making statements based on opinion; back them up with references or personal experience.


          Use MathJax to format equations. MathJax reference.


          To learn more, see our tips on writing great answers.




          draft saved


          draft discarded














          StackExchange.ready(
          function () {
          StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fdatascience.stackexchange.com%2fquestions%2f44595%2fwhy-is-the-keras-model-always-predicting-the-same-class-how-can-i-improve-the%23new-answer', 'question_page');
          }
          );

          Post as a guest















          Required, but never shown





















































          Required, but never shown














          Required, but never shown












          Required, but never shown







          Required, but never shown

































          Required, but never shown














          Required, but never shown












          Required, but never shown







          Required, but never shown







          Popular posts from this blog

          How to label and detect the document text images

          Vallis Paradisi

          Tabula Rosettana