Wafèlmolt

@superalex sorry for delay. I'm gonna investigate on my own instance; For now the only thing I'm sure is that the second POST request (for download trigger) is not sent...

Hi. Same issue for me on `devstral-small-latest` while trying to make it read a file.

Hello @zwpaper. I will try next monday. But I'm still wondering, why would chat feature require nvidia based image while code completion is working correctly ?

Oh ok. I guess there is a misunderstanding there : **My VSCode extension is running inside a devcontainer** whereas tabby server is actually running on my Ubuntu WSL (and clearly...

Good evening. I didn't try @zwpaper's solution yet. I will try tomorrow if I find time and keep you informed.

Hello. Confirmed what you suggested @zwpaper. It's working fine when using host network AND using localhost instead of bridge IP. Also I figured out that the issue is also happening...

> But does it mean while tabby completion is running in container, tabby chat is on the other hand running on VSCode host/front side ? When using host AND raw...

I edited my original issue's body to add some context on my environment.

> > Switched back to non-host bridge mode on my container and put "localhost" as alias of my bridge in /etc/hosts. And both completion and chat are working now. >...

Up 🖐