☆ Yσɠƚԋσʂ ☆@lemmygrad.ml to Technology@lemmygrad.mlEnglish · edit-24 days agoQwQ-32B is a 32 billion parameter language model achieves comparable performance to DeepSeek-R1 with 671 billion parameters, using reinforcement learning for scalingqwenlm.github.ioexternal-linkmessage-square3fedilinkarrow-up11arrow-down10cross-posted to: technology@lemmy.ml
arrow-up11arrow-down1external-linkQwQ-32B is a 32 billion parameter language model achieves comparable performance to DeepSeek-R1 with 671 billion parameters, using reinforcement learning for scalingqwenlm.github.io☆ Yσɠƚԋσʂ ☆@lemmygrad.ml to Technology@lemmygrad.mlEnglish · edit-24 days agomessage-square3fedilinkcross-posted to: technology@lemmy.ml
minus-squaremarl_karx@lemmygrad.mllinkfedilinkEnglisharrow-up0·2 days agoIsnt deepseek based on qwen? at least the distilled models?
minus-square☆ Yσɠƚԋσʂ ☆@lemmygrad.mlOPlinkfedilinkarrow-up0·2 days agoI think so, but this looks like an update of qwen with some new tricks.
Isnt deepseek based on qwen? at least the distilled models?
I think so, but this looks like an update of qwen with some new tricks.