I am working on a personal project on image classification (two classes) and am trying to see how the MobileNet v2 structure would perform. While training the training accuracy is already quite high after the first epoch after which is continues to increase. The validation accuracy, however, seems forever stuck around the 50% mark.

enter image description here

I have tried both SGD, Adam, and RMSprop as optimizers, however they both produce the same results. I have also used both a MobileNet v2 with and without ImageNet weights without seeing any difference. Other CNN architectures work perfectly fine, achieving up to ~95% validation accuracy. I am using the following code for my model:

from keras.applications import MobileNetV2
from keras.models import Sequential
from keras.layers import GlobalAveragePooling2D, Dropout, Dense

basemodel = MobileNetV2(input_shape=(128, 173, 3), include_top=False, weights=None)
model = Sequential(
        Dense(64, activation="relu"),
        Dense(1, activation="sigmoid"),
model.compile(loss="binary_crossentropy", optimizer="sgd", metrics=["accuracy"])

Is there something obvious that I am missing or am I just using incorrect hyperparameters for my dataset?

  • 1
    $\begingroup$ In the MobileNetV2's constructor, you need to set weights="imagenet". $\endgroup$ – Shubham Panchal Nov 29 '19 at 0:43
  • $\begingroup$ @ShubhamPanchal I also tried that, but it did not change anything with regards to the validation accuracy. $\endgroup$ – Oxbowerce Nov 30 '19 at 10:05

Well the critical step for finetuning/transfer learning any model in tensorflow[1,2+]:

| improve this answer | |
  1. Use the imagenet weights.
  2. Freeze the basemodel to prevent it from getting trained on your data.
  3. You might be overfitting the trainset.Try to not use the whole MobileNet model, maybe only the first 10-15 layers.
| improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.