Woo Kimm

Computer Vision

Results 3 comments of


                                            Woo Kimm

Any plan to release the cc3m filtered data and the weight of the linear layer after the 1st training stage?

+1

Training loss curve like stairs.

Hi! Have you solved the problem?

为何可以产生prompts

我尝试了 deepseek 放的 r 系列模型出 prompt 的概率很小，即使是他们蒸馏的 qwen 系列模型也是，你们有计划生成 r 系列模型的 prompts 吗？