82 2 weeks ago

A Deepseek-r1:7b Model with Deepseek-r1:8b temp and TopP, while having Deepseek-r1:1.5b as a speculative draft model

thinking