Open jiuntian opened 7 months ago
Hi,
Thanks for a great work in visualising the cross attention of diffusion models. Is it possible to visualise the cross-attention map for the tokens of <|startoftext|> and <|endoftext|>? Since they play an important role in diffusion models too.
Good point, I'll get around to adding that.
Hi,
Thanks for a great work in visualising the cross attention of diffusion models. Is it possible to visualise the cross-attention map for the tokens of <|startoftext|> and <|endoftext|>? Since they play an important role in diffusion models too.