echohive42 patreon

GPT resaercher Navigates Scientific Literature Code files

🕑 Added 2024-02-14 08:51:23 +0000 UTC

Comments

Echo Hive

I did processing of json responses before completed in the salesperson video. you might find it helpful: https://youtu.be/Cq0bYU93PGo

Michael Woodrum

I will play with this more when I get home. I am successfully streaming using yielding normally because of the async responses and it comes in and chunks and then I combine them and display them to the user as if it were typing across the front ends chat response. I will need to see what happens when streaming it to the front end as json. Then format once enough information exists to do so? I didn't know, I just need to test. I'm new to fastapi and it introduces many new ways of thinking compared to what I was doing in flask.

Echo Hive

You can use json mode all the time. but it feels more like a regular chat without json mode. json mode should work fine with the async versions but the streaming portion is just printing to the terminal and not yielding but returning the whole response when done within the method. just keep that in mind if you were trying to accomplish streaming responses with fastapi

Michael Woodrum

I have been working to get the unified open AI setup that you have made completely converted over for fast API and everything with fast API has to use asynchronous all the time and I'm having an issue with the JSON mode portion of it and I'm not exactly sure why. That being said why is it that you would not use JSON mode all the time?