Mr_Go_is_surfing
Mr_Go_is_surfing
> The fine-tuning part is not the focus of this paper, but we promise the results are reproducible and will make it public after the conference. If it's urgent for...
Hi! Sorry to reopen the issue again. Do you remember which setting is more accurate, just the cosine similarity or the Euclidean distance? Thanks!
Is there any mistake? What do you mean by "the Euclidean distance is better than Euclidean distance" haha
Thanks for your response! I also have another few questions: (1) how long does it take for each stage in DreamVideo? I have tried in my own server and found...
Thank you very much! By the way, how long does it take to evaluate on all datasets mentioned in the paper of DreamVideo? Could you provide the evaluation code?
Hi! Have you reproduced the group of experiment of cps_meanteacher_3b_w1.5_?
No, I haven't got the code yet. Could you successfully reproduce the results of Cityscapes?
@tpoisonooo hi，方便问下为什么muon用在sft不会好吗？目前有哪些实验支撑这一点的？
hello，请问你是怎么解决V100无法安装flash-attn导致安装不了openrlhf[vllm]的问题呢？