3 ms·
Right now I get 59 tok/sec on GPT-OSS 120B using Unsloth's dynamic 4-bit quants, via llama.cpp https://news.ycombinator.com/item?id=45881049 https://news.ycombi
by aseipp 11mo ago
Right now I get 59 tok/sec on GPT-OSS 120B using Unsloth's dynamic 4-bit quants, via llama.cpp https://news.ycombinator.com/item?id=45881049 https://news.ycombinator.com/item?id=45881049