Any one using autokeras on colab.research.google.com with gpu?

[Python-Ninja-Hebi]

Hello,

I used as example ImageClassifier.ipynb with gpu on colab.research.google.com.

In cell

clf.fit(x_train, y_train, epochs=3) # each model is trained for three epochs

I got

UnknownError Traceback (most recent call last)
in ()
7 # 20% of the data is used as validation data by default for tuning
8 # the process may run for a bit long time, please try to use GPU
----> 9 clf.fit(x_train, y_train, epochs=3) # each model is trained for three epochs

15 frames
/usr/local/lib/python3.7/dist-packages/tensorflow/python/eager/execute.py in quick_execute(op_name, num_outputs, inputs, attrs, ctx, name)
58 ctx.ensure_initialized()
59 tensors = pywrap_tfe.TFE_Py_Execute(ctx._handle, device_name, op_name,
---> 60 inputs, attrs, num_outputs)
61 except core._NotOkStatusException as e:
62 if name is not None:

UnknownError: Failed to get convolution algorithm. This is probably because cuDNN failed to initialize, so try looking to see if a warning log message was printed above.
[[node model/conv2d/Conv2D (defined at /usr/local/lib/python3.7/dist-packages/autokeras/utils/utils.py:88) ]] [Op:__inference_train_function_9811]

Function call stack:
train_function

What can I do?

Best regards

Hebi