celestialxevermore

It is excactly what I've expected too

Thank you for replying. As I know, The temporal encoder, Transformer is randomly initialized, which causes some sub-optimal phenomenon as the randomly initialized weights of the seqTransf do harm on...

Oh, Really Thank you for your very kind and fast reply. I didn't notice that [line 116 of modules/modeling.py](https://github.com/xuguohai/X-CLIP/blob/6b5344f44537d758acb82d115b8484f7430f9fb0/modules/modeling.py#L116) means that code makes initialized the seqTransf from clip. Q1. Then,...