Rasa Training Error: "Invalid Instruction" on Linux Debian - Need Help!

Hello everyone,

I’ve been struggling with a frustrating issue while trying to set up and train a Rasa chatbot on my Linux Debian system, and I’m hoping someone here can help me out. Here’s the full story:

My Setup:

  • OS: Linux Debian
  • Python Version: 3.9.x (I’ve tried updating it multiple times)
  • Rasa Version: Tried different versions, including the latest
  • Hardware: Pretty old PC with weak specs, not much RAM or processing power

The Issue:

Whenever I try to train the Rasa model using the rasa train command, I keep running into an “Invalid Instruction” error. This happens even when I just create a new project with rasa init and use the default files. Here’s a snapshot of what I see:

text

Копировать код

ERROR: Command errored out with exit status 1:
...
2024-07-29 14:56:57 DEBUG rasa.shared.nlu.training_data.loading - Training data format of 'data/stories.yml' is 'unk'.
Invalid instruction.

This error has been driving me crazy for days now. I’ve tried a bunch of things, including:

  1. Creating and activating a new virtual environment: Didn’t solve the issue.
  2. Reinstalling Rasa and dependencies: I’ve tried different versions, including locking sqlalchemy to <2.0.
  3. Upgrading setuptools, pip, and other packages: Still no luck.
  4. Debug Mode (rasa train --debug): Just shows the same “Invalid Instruction” error without much more detail.
  5. Manually installing specific versions of dependencies like pymongo and sqlalchemy.
  6. Checking system logs and pip check: Nothing unusual found.

No matter what I do, the error persists. It’s at the training stage that things go wrong, and I have no idea why.

I’m really stuck and could use some fresh eyes on this. If anyone has faced a similar issue or has any idea what might be going wrong, please let me know. My system is pretty old, so I wonder if that’s a factor, but I can’t easily switch to a different machine.

Any help or suggestions would be greatly appreciated. Thanks in advance!