Using Q38FN with dflash2? Try lower draft model quants.
No snapshot captured; read the original:
https://www.reddit.com/r/LocalLLaMA/comments/1xgozlq/using_q38fn_with_dflash2_try_lower_draft_model/No snapshot captured; read the original:
https://www.reddit.com/r/LocalLLaMA/comments/1xgozlq/using_q38fn_with_dflash2_try_lower_draft_model/