Issues with training rasa model on redhat

hi… I’ve built a chatbot with rasa and it’s currently deployed in my ubuntu 18.04 with nginx. But now I have to host it on a server computer which is redhat 7 version of linux. I’ve tried searching for installation guides on redhat… I could not find any… so I followed the ubuntu installation guide which led me into problems… whenever I’m trying to train my model it’s saying segmentation fault…

2020-02-19 16:33:18 INFO rasa.model - Data (stories) for Core model section changed. Training Core model… 2020-02-19 16:33:24 INFO absl - Entry Point [tensor2tensor.envs.tic_tac_toe_env:TicTacToeEnv] registered with id [T2TEnv-TicTacToeEnv-v0] Processed Story Blocks: 98%|██▉| 41/42 [00:01<00:00, 22.21it/s, # trackers=235]Segmentation fault

But I’ve trained this model with no problem in my venv in ubuntu… can anyone help me with anything related to installing rasa on a redhat machine! Thanks!

Hi @AnikaTabassum! This is a reported issue that is fixed but not released yet. However, there is a fix:

pip install gym==0.15.4
1 Like