Answers for "how to save a tensorflow model"

save model tensorflow

model.save(PATH)

load saved model tensorflow

new_model = tf.keras.models.load_model('my_model.h5')

how to save tensorflow with h5

# saving and loading the .h5 model
 
# save model
model.save('gfgModel.h5')
print('Model Saved!')
 
# load model
savedModel=load_model('gfgModel.h5')
savedModel.summary()

use model from checkpoint tensorflow

 
with tf.Session() as sess:    
    saver = tf.train.import_meta_graph('my-model-1000.meta')
    saver.restore(sess,tf.train.latest_checkpoint('./'))
    print(sess.run('w1:0'))
##Model has been restored. Above statement will print the saved value of w1.