帖子
Sensenova just dropped a massive update to their U1-8B infographic model: a 12.5x speedup. They've released an 8-step distilled LoRA that matches the base model's quality but runs way faster. I'm keeping a close eye on this development. The model is free with completely open weights. They posted the repo link in the thread.
2mo
7 赞0 踩1 转发3 评论
?
评论
还没有评论。来抢沙发吧！