Open Jonii opened 2 weeks ago
Probably and issue with the cudnn
specific implementation on the tf backend, which is pretty dense. I will take a look.
Similar issue happening in case of running Keras with Tensorflow backend on desktop. Hidden states of the individual element in a batch are returned as a tuple of the GRU output and not as a Tensor with first dimension equal to batch size.
I opened this on tensorflow repo, and was told to move it here: https://github.com/tensorflow/tensorflow/issues/74475
The short of it, gru, at least on google colab(keras 3.4.1) returns wrong things when run with gpu available.
Minimal way to reproduce here:
This prints
With GPU:
With CPU:
CPU behavior seems correct.
**Edited to add, I do not have the ability to test gpu behavior outside of google colab, so this might be a bug that's been fixed on the latest version, or due to colab-specific misconfiguration.