I tried to used KNIME with conda environment. I tested the workflow from this tutorial with the workflow.
I install a new conda env containing:
tensorboard 2.10.0 py38haa95532_0
tensorboard-data-server 0.6.1 py38haa95532_0
tensorboard-plugin-wit 1.8.1 py38haa95532_0
tensorflow 2.10.0 mkl_py38ha5c4042_0
tensorflow-base 2.10.0 mkl_py38h6a7f48e_0
tensorflow-estimator 2.10.0 py38haa95532_0
keras 2.10.0 py38haa95532_0
keras-preprocessing 1.1.2 pyhd3eb1b0_0
pandas 2.0.3 pypi_0 pypi
When running, it threw an error:
knime_5.2.1\plugins\org.knime.dl.python_5.2.0.v202310301554\py\DLPythonNetwork.py", line 53, in <module>
from pandas.util.testing import network
ModuleNotFoundError: No module named 'pandas.util.testing'
I tried to read the solution in this discussion. However, I think that the solution did not solve the problem because keras code are using old version of pandas. I would like to update the code to match with later version of pandas. Please help me to solve it. Thank you very much.
Update1: Install pandas version 1.5.3
It threw a new error:
.conda\envs\py38_knime\lib\site-packages\keras\src\engine\keras_tensor.py", line 270, in __hash__
raise TypeError(
TypeError: Tensors are unhashable (this tensor: KerasTensor(type_spec=TensorSpec(shape=(None, 4), dtype=tf.float32, name='input_1'), name='input_1', description="created by layer 'input_1'")). Instead, use tensor.ref() as the key.
Like this disscussion.
