TextVectorization: Missing support for saving to .h5
extreme4all opened this issue · 1 comments
extreme4all commented
save_format='tf' gives a folder and is less portable than an .h5
Please support saving to .h5
VOCAB_SIZE = 20_000
encoder = tf.keras.layers.TextVectorization(
max_tokens=VOCAB_SIZE,
standardize='lower',
output_mode='int')
encoder.adapt(train_dataset.map(lambda text, label: text))
The error:
NotImplementedError: Save or restore weights that is not an instance of tf.Variable
is not supported in h5, use save_format='tf'
instead. Received a model or layer TextVectorization with weights [<keras.layers.preprocessing.index_lookup.VocabWeightHandler object at 0x0000025548A6C6A0>]
broken commented
This layer is a part of Keras and not TF Text. Can you add this as an issue to it instead of here? Thanks.