ParentFull threadmhjkl·If you’re on AMD, use llamacpp with the vulkan backend. For the 35B I get around 115 tps without mtp and 150-170tps with it. Also the pcie slot you use affects itView on HN