Frontend microservice not running in "Building RAG Agents with LLMs" and other issues

AI & Data Science AI Foundation Models and Endpoints
1

In DLI course “Building RAG Agents with LLMs”, the frontend microservice is not activated at the start of the course environment. Because of this, I’m not able to complete the course assignment.
The issue is evident to me by running notebook 01_microservices.ipynb. This notebook prints the microservices that are running in the course environment. Until not long ago (when I started the course), it was printing also the frontend microservice as expected. It doesn’t print it anymore.

A few more issues:

  1. there seems to be a compatibility issue between the version of httpx installed and langserve (there is another topic that was recently opened about this). This doesn’t let me make progress with the course assignment.
  2. there are multiple versions of the instructions on how to complete the assignment and they either contradict each other or I don’t see how to combine them. 1 version is in the last video of the course. 1 version is in notebook 8. 1 version is in notebook 9. These versions regard different files and notebooks as crucial for successfully completing the assignment. Some of these files and notebooks no longer exist or have been renamed. It would be great if one could sort this out.

Thank you.

2

Hi @lorenzo.rossi299 thanks for the detailed information. I’ll pass this on to the right team now and see if we can get you sorted asap!

3

Hey @lorenzo.rossi299

Thanks for the notice! You were spot-on!

  • Despite most of packages being version-locked, the httpx moved and caused an issue on rebuild. I’ve updated the specs, and the issue seems to be resolved now. Please let me know if you run into something.
  • I checked over the notebooks and didn’t see much of a discrepancy between 8 and 9. Happy to clear up anything in a follow-up though. On the videos… yeah, I’d like to rerecord them at some point. The material has moved a couple of times since the initial recording, and there an effort to update this material a little over the next few weeks as well. I should have some time post-GTC, so that’s on the todo list. Thanks for the CTA though!
4

Thank you @vkudlay
I was able to complete the course.

In case it helps, I’ve identified my main cause of confusion about the instructions on how to complete the assignment. It’s the two pieces of text that I paste at the end of this message.
Combining those with the fact that notebook 9 wasn’t working (due to the issue that you fixed with the frontend), I started to think that I had to do something with the Python code for the frontend (maybe modify it or running it myself). Perhaps this confusion only arose in my case so feel free to ignore all of this.

Instructions that confused me

From notebook 8:

On launch, frontend/frontend_block.py had several lines of code which trigger the course pass condition. Your objective is to invoke that series of commands by using your pipeline to pass the Evaluation check!

From notebook 9:

Using this, you will be able to supply functionality to the frontend service’s frontend_server.py session, which strongly expects:

  • A simple endpoint named :9012/basic_chat for the basic chatbot, exemplified below.

  • A pair of endpoints named :9012/retriever and :9012/generator for the RAG chatbot.

  • All three for the Evaluate utility, which will be required for the final assessment.

1 Like
5

Awesome! Thanks so much! I’ll think about if there’s a better way to say that in the next update :D

And congrats! Glad you finished