Zhang Wenbin
Zhang Wenbin
同样问题，我看过data_process.py 输入512 * 512 实际默认 450*450尺寸，除了将生成的视频再进行分辨率增加之外，能否直接支持输入更大分辨率的视频？ @ashawkey 希望能够得到解答，感谢！
我看论文上写512*512 25帧 视频 15秒训练
https://huggingface.co/THUDM/codegeex2-6b
@wac81 改哪个参数？
已支持 @hsiunien
遇到类似问题，有谁知道如何解决吗？
我训练25个迭代是正常的，Irs3训练需要的文件在此目录下，如下图所示：
I tested Qwen/Qwen2.5-7B-Instruct-1M using DualChunkFlashAttention backend. It startup well, but not work well. @sighingnow ubuntu-vllm-openai-1 | INFO 05-31 19:13:07 [logger.py:42] Received request cmpl-77d91882816c4f748e2023c93449f62d-0: prompt: 'Once upon a time', params: SamplingParams(n=1,...
> load_checkpoint_and_dispatch It works!
> > > Hello, [@mxchinegod](https://github.com/mxchinegod). Thank you for pointing this issue! We've released the latest version 1.3.4 on PyPi, where this issue has been fixed. > > > > >...