Difference between eva-clip model and MIM pretrained model

[jihaonew]

I find that the eva-clip model has an extra inner_attn_ln layer compared to the original pretrained model.

EVA-CLIP

EVA

Any reason to use another layer-norm in the clip model?

[Owen1u]

您好，我已收到您的邮件，我会尽快给您回复。

[moon-wh]

Have you got answer yet? I have the same question.