We are running Rasa on an Ubuntu 22.04 (AWS EC2 server). This is for a personal project that I’m working on. I see that, when I hit ‘rasa train’,the training starts and at around 66%, I see the ‘# trackers’ is at 16384 and it’s eating up all the memory. From this point on wards, the memory consumption starts to spike and the training percentage doesn’t move forward. I need tips to impose memory consumption limits. Run the training with some memory boundaries. Please help me to understand these trackers that I see in the status. What can I do to bring that number down? I’ve attached a screenshot of the status that I see. [Screenshot 2024-08-03 at 10.07.05 AM]

Training Rasa on an Ubuntu 22.04 eats up all the memory

cravi47 (Raviteja) August 3, 2024, 3:27pm 2

Update: The number of trackers that I mentioned, just increased to 1048576. Is there a way to limit them? What are these trackers?

I’m a newbie to Rasa.

Topic		Replies	Views
Memory issue in training Getting Started with Rasa	0	130	December 20, 2018
Reducing or limiting memory usage during rasa train Rasa Open Source	4	1302	March 16, 2021
RASA Training Memory Issues Rasa Open Source	2	848	December 31, 2019
How can I give Rasa more resources for training? Rasa Open Source	2	831	July 1, 2020
Any way to to check Training process Rasa Open Source	2	695	July 16, 2019