search

brianpetro / obsidian-smart-connections

Chat with your notes & see links to related content with AI embeddings. Use local models or 100+ via APIs like Claude, Gemini, ChatGPT & Llama 3
https://smartconnections.app
GNU General Public License v3.0
2.84k stars 185 forks source link

Error first time loading Smart Connections v 2.2.76 #799

Open AlexCuan opened 2 months ago

AlexCuan commented 2 months ago

image

Tried a clean reinstall and nothing. My OS is Manjaro 24.0.8

brianpetro commented 2 months ago

Seems to be because a markdown file extension is capitalized.

It should be case-insensitive, so that's something I'll have to fix.

Making that file extension lowercase should fix it in the meantime, though 🌴

AlexCuan commented 2 months ago

You were right, after fixing the extension of the files error dissapeared. Now there is this one:

image

Do you know what could be the fix? I have an RTX 2060 and can use it normally, so there is no gpu missing problem

Below there's another one that says:

image

brianpetro commented 2 months ago

@AlexCuan thanks for the update.

Local embedding models should work if you toggle on the "legacy transformers" setting.

Unfortunately the GPU flag would have to be turned on by Obsidian to get the GPU to work within the plugin on a Linux machine.

However, the next Smart Connect app update will re-introduce embedding offloading from Obsidian to the Smart Connect app. The Smart Connect app has the necessary flag for enabling GPU, so you will be able to run that in the background and it will handle all the embeddings via GPU even on Linux.

🌴

AlexCuan commented 2 months ago

Perfect! I'm really looking forward to the next update 👏🏻. Using the dedicated GPU will speed things up a lot. Thanks so much for your help! 🙏🏻