search

lmoroney / dlaicourse

Notebooks for learning deep learning
5.66k stars 5.36k forks source link

how to save and load models that was trained using fit_generator #45

Closed sleekmike closed 4 years ago

sleekmike commented 4 years ago

Hello, I have been trying to reproduce the code(human or horse classification) in tensor flow in practise course (/Course 1 - Part 8 - Lesson 4 - Notebook.ipynb ) locally on my pc. It works so far but when I save a model with this line:

history = model.fit_generator( train_generator, steps_per_epoch=8,
epochs=15, verbose=1, validation_data = validation_generator, validation_steps=8)

try: model.save_weights('human_horse1.h5') except Exception as e: print("error in saving model: ", e)

and to load the saved: new_model = tf.keras.models.load_model('human_horse3.h5')

if I try to used the saved model to predict with the code:

new_model.predict('./Users/macbook/Desktop/image_classifier/human01-04.png')

I normally get this error: AttributeError: 'str' object has no attribute 'shape'

please how do I properly load and save models and use the model to predict locally on my pc. Kind regards.

JonathanSum commented 4 years ago

I am not really sure this can solve your issue. "shape" attribute is from numpy. Have you installed numpy yet?

sleekmike commented 4 years ago

Yes I have numpy and tensorflow 2.0 installed on my system.

On Fri, 4 Oct 2019 at 1:36 PM, Jonathan Sum notifications@github.com wrote:

I am not really sure this can solve your issue. "shape" attribute is from numpy. Have you installed numpy yet?

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/lmoroney/dlaicourse/issues/45?email_source=notifications&email_token=AI5IK7A5J4XAZDLEEJG6CT3QM42GPA5CNFSM4I5CNMUKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEALQDTY#issuecomment-538378703, or mute the thread https://github.com/notifications/unsubscribe-auth/AI5IK7EJRYS5RM7XFM4ND73QM42GPANCNFSM4I5CNMUA .

cicimeng commented 4 years ago 
    why i can't install tensorflow 2.0 upgrade doesn't work 

    Get Outlook for iOS

On Fri, Oct 4, 2019 at 9:40 PM +0800, "Mike Ohanu" notifications@github.com wrote:

Yes I have numpy and tensorflow 2.0 installed on my system.

On Fri, 4 Oct 2019 at 1:36 PM, Jonathan Sum notifications@github.com

wrote:

I am not really sure this can solve your issue. "shape" attribute is from

numpy. Have you installed numpy yet?

—

You are receiving this because you authored the thread.

Reply to this email directly, view it on GitHub

https://github.com/lmoroney/dlaicourse/issues/45?email_source=notifications&email_token=AI5IK7A5J4XAZDLEEJG6CT3QM42GPA5CNFSM4I5CNMUKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEALQDTY#issuecomment-538378703,

or mute the thread

https://github.com/notifications/unsubscribe-auth/AI5IK7EJRYS5RM7XFM4ND73QM42GPANCNFSM4I5CNMUA

.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub, or mute the thread.

JonathanSum commented 4 years ago

I think the problem is this code: new_model.predict('./Users/macbook/Desktop/image_classifier/human01-04.png'), and you should remove it.

And add these codes and use them: import numpy as np from keras.preprocessing import image

img_width, img_height = 150, 150 img = image.load_img('./Users/macbook/Desktop/image_classifier/human01-04.png', target_size = (img_width, img_height)) img = image.img_to_array(img) img = np.expand_dims(img, axis = 0) model.predict(img) If that does work, I hope someone who can come here and help you. Or you can show us your code, and I will try to fix it with my cpu only laptop this week.

sleekmike commented 4 years ago

Thanks a lot @JonathanSum, I used your code and it works like magic :).