Open vschram opened 1 month ago
Thanks for the report!
I expect the issue is that you are using an old (pre 1.0) version of trieste. Are you able to update to a more recent one?
There may also be an issue with the fact that the current version of GPflow doesn't fully work with TensorFlow 2.16 due to breaking changes in the latter (though the upcoming GPflow release will). You can work around this either by downgrading to TensorFlow 2.15, or by following the suggestion in this comment.
Dear Secondmind Team,
Describe the bug When I am trying to reproduce the Trieste tutorials, e.g. : https://secondmind-labs.github.io/trieste/3.1.0/notebooks/expected_improvement.html or https://secondmind-labs.github.io/trieste/3.1.0/notebooks/inequality_constraints.html I am getting the following error:
\Anaconda3\envs\clone_BayesOpt_env\Lib\site-packages\tensorflow\python\module\module.py", line 428, in _flatten_module raise ValueError("Error processing property {!r} of {!r}".format( ValueError: Error processing property 'data' of _TupleWrapper((<tf.Variable 'Variable:0' shape=(None, 2) dtype=float64, numpy= array([[4.05741721, 0.49318051], [5.55239509, 1.09468531], [1.60541616, 5.99931802], [0.37097807, 4.92991607], [2.9573046 , 4.75401206]])>, <tf.Variable 'Variable:0' shape=(None, 1) dtype=float64, numpy= array([[-1.02023562], [-0.61750123], [ 0.0417215 ], [ 0.52161525], [ 0.22206321]])>))
To reproduce Steps to reproduce the behaviour:
Expected behaviour Expected to see: Optimization completed without errors
System information
Thanks in advance for any hint.