How the shape of integer sequences is determined in TextVectorization preprocessing layer? and how to handle the shape?
Can you post the code that got you to this output?
inp_text_preprocessor = ???
inp = ???
I think mostly you’re just looking for the “output_sequence_length” argument.
https://www.tensorflow.org/api_docs/python/tf/keras/layers/TextVectorization
If none, it’s padded to the longest sequence, otherwise it’s padded or truncated to fit.
1 Like
Yes, @markdaoust I’m confused with the output_sequence_length. How did it come 64 in the first pic?
So if I specify output_sequence_length=None it will automatically get padded to the longest sequence?
specify output_sequence_length=None it will automatically get padded to the longest sequence?
It’s supposed to.
input is a list of Portuguese sentences.
This screenshot doesn’t help anything, most of the arguments are not shown.
1 Like
Thank you @markdaoust I think I got the point we can control the output sequence length with the parameter output_sequence_length.