I am training a model using keras, and it trains great. I get clear results, and i can even save it with model.save('saved_model'). However when I try to load the 20mb model again with load_model('saved_model'), it stalls, eating up one of my processors at 100%, but even after a half hour of waiting it does not load. Here is my code:
callbacks = ModelCheckpoint('_checkpoint',
monitor='val_loss',
verbose=1,
save_best_only=True,
mode='min')
model = _keras_compiled_model(x_dim,y_dim,len(labels))
history = model.fit(x_train, y_train,
batch_size=batch_size,
epochs=epochs,
validation_data=(x_test, y_test),
callbacks=[callbacks])
# load the model with the best validation loss, delete checkpount data
model.load_weights('_checkpoint')
model.save('saved_model')
What could I be doing wrong here? keras, Tensorflow 2.0, 1080ti, Windows 10