logoalt Hacker News

vineyardmike • today at 7:57 PM • 1 reply • view on HN

According to TM, the creators, they benchmarked it to run at 5x the speed of Qwen 27, so there is probably some issue on your end. It runs a fraction of the active parameters, even with looping, so it should be much faster.


Replies

syntaxing • today at 7:58 PM

This has been observed by the broad community for llama cpp.

➕ show 1 reply