How can we plot accuracy and loss graphs from a Keras model saved earlier?

Question

Is there a way to plot accuracy and loss graphs from the CNN model saved earlier? Or can we only plot graphs during training and evaluating the model?

model          


        

Answer 1

It depends on the way you saved the model.

In general, there are two cases, the first one is saving and loading the whole model (including architecture and weights):

from keras.models import load_model

model.save('my_model.h5')  # creates a HDF5 file 'my_model.h5'
...
model = load_model('my_model.h5')

The second is saving only the weights:

def create_model():
   model = Sequential() 
   # ... creating the model exactly as it was defined in the training time

# You must create the model since loading only the weights
model = create_model()
model.load_weights('my_model_weights.h5')

For more details read Keras documentation