Orca TF2 estimator cannot load tf savedmodel directly

[cyita]

I load tf savedmodel using

def model_creator(config):
    item_model = keras.models.load_model("/opt/work/guoqiong/friesian_embed/Training/Embedding/2tower_models/exp5/item-model", custom_objects={'tf':tf})
    optimizer = tf.keras.optimizers.Adam(config["lr"])
    item_model.compile(optimizer=optimizer, loss='binary_crossentropy', metrics=['binary_accuracy', 'binary_crossentropy', 'AUC'])
    return item_model

est = Estimator.from_keras(model_creator=model_creator, verbose=False, config=config)

And every node should have a copy of this savedmodel. Orca TF2 estimator may add a function that loads tf savedmodel directly.

[songhappy]

an idea, can you try to load the model locally, and put it into config? config["mode"] = keras.models.load_model("/opt/work/guoqiong/friesian_embed/Training/Embedding/2tower_models/exp5/item-model", custom_objects={'tf':tf}) def model_creator(config): item_model = config["model"] return item_model