Very impressed by the Vulkan API support

[grondilu]

I have a gaming laptop with a RADEON GPU and I had never managed to make it work for AI inference before.

Here it works great, I get blazing fast output from Llama3.

Well done.

[enema-combatant]

Same here! Within 10 minutes I was up and running, and I'm slow! I have a 2080 Super, and it produces nearly as fast as the major cloud AIs. Stunning.