generate_and_save_report() crashing the Colab kernel

[achahid]

Has anyone run into eda.silently (no traceback, just a restart)? I've isolated it to the _generate_relationship_among_variables_card step — the other cards generate fine individually, and the crash happens whether or not the model is fitted. Is this a known issue, and is there a recommended workaround?

[dattatreyam23]

Hi @achahid ,

Thank you for contacting us!

We looked into this and tested the exact same code, which runs without any issues in our Getting Started Colab.

Based on your setup details, the only difference we noticed is that you are using TensorFlow 2.19, whereas the default Colab runtime uses TensorFlow 2.20.

To troubleshoot this, we recommend the following steps:

• Upgrade TensorFlow: Please upgrade your environment to TensorFlow 2.20 and rerun your code. This version mismatch is the most likely culprit.

• Isolate the data: If upgrading doesn't resolve the issue, please try running the Getting Started Colab in your current setup. If the Colab notebook works perfectly for you there, the issue is likely stemming from the specific data you are using rather than the code itself.

Let us know if upgrading to TF 2.20 resolves the issue, or if you need further help investigating!

Thank you,

Google Meridian Support Team