François Chollet
François Chollet
Why do you use `tf.function` though?
Just like in layers, you're supposed to override `call()`, not `__call__()`. You could override `__call__()`, but by doing so you miss a bit of built-in functionality, including auto-broadcasting. So you...
Your custom type should include everything you want to be recognized as a tensor, which is dependent on your use case. For instance, it could be the union of tf.Tensor,...
The lowercasing is simply does via the TensorFlow operation `tf.strings.lower`, and since it needs to be a TF op, we are not at liberty to change it. You could open...
`x.shape` is the "static shape" of `x`. It is often not a number. It can be `None`. If you want the actual number value, you must use `keras.ops.shape(x)`, e.g. ```python...
There are no immediate plans, but Keras will likely add support once it is robustly available in torch, TF, and JAX.
Sure, you are welcome to work on that. Do you have any experience with web development? I'm thinking this tool may benefit from an interactive js/html interface to be used...
Sounds great. Here's an example of a draft I wrote a long time ago, displaying a kind of summary of a file's content: ```python def inspect_file( filepath, reference_model=None, custom_objects=None, print_fn=print...
Sure, you can just ask questions in this thread.
> Saving the same model as .h5 works (unfortunately this workaround is not suitable for us because our technical requirement is to save a .pb-model). Is it because you want...