i am still new to deep learning right now i follow this keras tutorial to make an translation model using transformer here the link. Everything works just fine but i have no idea how to save the model, currently i have this implementation
<code>transformer.save("models/transformer.h5")
</code>
<code>transformer.save("models/transformer.h5")
</code>
transformer.save("models/transformer.h5")
and load the model with this
<code>new_model = tf.keras.models.load_model(
"models/transformer.h5",
custom_objects={
"PositionalEmbedding": PositionalEmbedding,
"TransformerEncoder": TransformerEncoder,
"TransformerDecoder": TransformerDecoder,
},
)
new_model.summary()
</code>
<code>new_model = tf.keras.models.load_model(
"models/transformer.h5",
custom_objects={
"PositionalEmbedding": PositionalEmbedding,
"TransformerEncoder": TransformerEncoder,
"TransformerDecoder": TransformerDecoder,
},
)
new_model.summary()
</code>
new_model = tf.keras.models.load_model(
"models/transformer.h5",
custom_objects={
"PositionalEmbedding": PositionalEmbedding,
"TransformerEncoder": TransformerEncoder,
"TransformerDecoder": TransformerDecoder,
},
)
new_model.summary()
i also attempt to save the vectorization and load it with pickle following this
but when i use the model outside the notebook the model is worsen, when i load the model in the same notebook it just fine (without reloading the vectorization)
can someone guide me how to correctly how to save this model?