bug: [API] chat completions

[LeGitHubDeTai]

Steps to reproduce
Make an API request with the stream mode set to true.

Expected behavior
When I make a request to v1/chat/completions with the stream mode set to true, everything works correctly. However, when it is set to false, the computer becomes overloaded, and no response is received from the API (worked in 0.4.4).

Environment details
Operating System: Windows 11
Jan Version: 0.4.5
Processor: Intel Core i9
RAM: 64GB
Any additional relevant hardware specifics: RTX 3090

[Van-QA]

hi @LeGitHubDeTai, sorry for the inconvenience. The issue may be related to this janhq/cortex#391. which we are still working on it.

[LeGitHubDeTai]

ow my bad sorry

[Van-QA]

hi @LeGitHubDeTai, we resolved the issue, and I tested both UI and the curl with Stream = false ✅ Jan v0.4.5-215

Feel free to reopen this if the issue persists.

[LeGitHubDeTai]

thanks !