google colaboratory, weight download (export saved models)

Question

I created a model using Keras library and saved the model as .json and its weights with .h5 extension. How can I download this onto my local machine?

to save the model I followed this link

Answer 1

This worked for me !! Use PyDrive API

!pip install -U -q PyDrive
from pydrive.auth import GoogleAuth
from pydrive.drive import GoogleDrive
from google.colab import auth
from oauth2client.client import GoogleCredentials

# 1. Authenticate and create the PyDrive client.
auth.authenticate_user()
gauth = GoogleAuth()
gauth.credentials = GoogleCredentials.get_application_default()
drive = GoogleDrive(gauth)

# 2. Save Keras Model or weights on google drive

# create on Colab directory
model.save('model.h5')    
model_file = drive.CreateFile({'title' : 'model.h5'})
model_file.SetContentFile('model.h5')
model_file.Upload()

# download to google drive
drive.CreateFile({'id': model_file.get('id')})

Same for weights

model.save_weights('model_weights.h5')
weights_file = drive.CreateFile({'title' : 'model_weights.h5'})
weights_file.SetContentFile('model_weights.h5')
weights_file.Upload()
drive.CreateFile({'id': weights_file.get('id')})

Now, check your google drive.

On next run, try reloading the weights

# 3. reload weights from google drive into the model

# use (get shareable link) to get file id
last_weight_file = drive.CreateFile({'id': '1sj...'}) 
last_weight_file.GetContentFile('last_weights.mat')
model.load_weights('last_weights.mat')

A Better NEW way to do it (post update) ... forget the previous (also works)

# Load the Drive helper and mount
from google.colab import drive
drive.mount('/content/drive')

You will be prompted for authorization Go to this URL in a browser: something like : accounts.google.com/o/oauth2/auth?client_id=.....

obtain the auth code from the link, paste your authorization code in the space

Then you can use drive normally as your own disk

Save weights or even the full model directly

model.save_weights('my_model_weights.h5')
model.save('my_model.h5')

Even a Better way, use call backs, which automatically checks if the model at each epoch achieved better than the best saved one and save the one with best validation loss so far.

my_callbacks = [
    EarlyStopping(patience=4, verbose=1),
    ReduceLROnPlateau(factor=0.1, patience=3, min_lr=0.00001, verbose=1),
    ModelCheckpoint(filepath = filePath + 'my_model.h5', 
    verbose=1, save_best_only=True, save_weights_only=False) 
    ]

And use the call back in the model.fit

model.fit_generator(generator = train_generator,  
                    epochs = 10,
                    verbose = 1,
                    validation_data = vald_generator,
                    callbacks = my_callbacks)

You can load it later, even with a previous user defined loss function

from keras.models import load_model
model = load_model(filePath + 'my_model.h5', 
        custom_objects={'loss':balanced_cross_entropy(0.20)})