Rasa crashes when training on machine with no GPU

Rasa Open Source
1

Hello, I am running “train” from rasa-x front-end, using docker, on a machine with no GPU, and it crashes. Below is the last error message I get before the rasa worker restarts. My machine has no GPU, and hence in theory I should be able to “ignore” this error message, but still it seems like it causes the crash.

Can anyone help?

rasa-worker_1 | 2021-01-18 19:00:05.021683: W tensorflow/stream_executor/platform/default/dso_loader.cc:59] Could not load dynamic library ‘libcudart.so.10.1’; dlerror: libcudart.so.10.1: cannot open shared object file: No such file or directory rasa-worker_1 | 2021-01-18 19:00:05.021744: I tensorflow/stream_executor/cuda/cudart_stub.cc:29] Ignore above cudart dlerror if you do not have a GPU set up on your machine.

2

Never mind, I figured it out. I ran out of memory on a 16G machine, because I was using word a word embedding file of 7G