Running DeepSeek AI with .NET Aspire | Full-Stack FridAI

[@davebrock7356]

Excellent stuff as always!

[@TimeStampBuddy]

0:00:00 - Intro & Setup 0:14:45 - AI Integration 0:35:50 - Custom UI & AI 0:47:36 - Prompt Issues 1:02:38 - Model Differences 1:24:41 - Conclusion

[@jonathanl2580]

Cool, now provide that in the following json format.

[@TimeStampBuddy]

@@jonathanl2580 [ { "timestamp": "0:00:00", "title": "Intro & Setup" }, { "timestamp": "0:14:45", "title": "AI Integration" }, { "timestamp": "0:35:50", "title": "Custom UI & AI" }, { "timestamp": "0:47:36", "title": "Prompt Issues" }, { "timestamp": "1:02:38", "title": "Model Differences" }, { "timestamp": "1:24:41", "title": "Conclusion" } ]

[@EdwardCharbeneau]

Note: Some of the Prompt Issues were because I didn't add the response_json switch to the request. This works on most models including Deepseek, but I missed it in the API docs.

[@NotInventedHereShow]

Hey Ed, you want to delete the .vs folder to empty the vs cache and get things back working.

[@stevenrichardson7876]

For some reason mine is not using the GPU (4090), must be setting in docker ?

[@EdwardCharbeneau]

1) Try updating docker, this was an issue for me once. 2) Make sure the checkbox, Use the WSL 2 based engine, is enabled. 3) Update your GPU drivers (always update GPU drivers)