rasa in docker - tensorflow error "invalid opcode"

[trstn70]

I'm using this repository: https://github.com/sasastarcevic/rasa-core-and-nlu-docker-compose

Rasa NLU version:
The Repo uses image tag "latest" as it is not defined. The Dockerfile executes "pip install rasa_core". This seems to be the newest Version as of today.

Operating system:
Debian 9 VM on Proxmox

Issue:
the rasa container is being perodically restartet and i see a tensionflow error in the logs:

Jan 23 10:53:39 lxvm165-113 kernel: [ 7357.922165] traps: python[20676] trap invalid opcode ip:7f5368d55250 sp:7ffcd9265c20 error:0
Jan 23 10:53:39 lxvm165-113 kernel: [ 7357.922193]  in libtensorflow_framework.so[7f53688c6000+c88000]

[akelad]

Thanks for raising this issue, @MetcalfeTom will get back to you about it soon.

[MetcalfeTom]

Hmm this one's a little mysterious, I've been doing some researching and it seems to be due to an incompatibility between the OS and the tensorflow binary

As a first pass at a solution, could you try again using a different tensorflow binary (or perhaps even by compiling one from source) and get back to us with your results

[no-response]

This issue has been automatically closed because there has been no response to our request for more information from the original author. Without this, we don't have enough information to help you. Please comment below with the requested information if you still need help.