4 ms·
From what I can find online LLAMA-65B 4-bit quantized can run 1 token/s on a Ryzen 7 3700X (using llama.cpp).
by GaggiX 3y ago
From what I can find online LLAMA-65B 4-bit quantized can run 1 token/s on a Ryzen 7 3700X (using llama.cpp).
loading thread…