Question about Figure 5 in the paper

[yanzengatsmu]

I guess the graph in Figure 5 represents the changing of cosine similarity between x_t and y_t. From left to right, the first part(50 to 0) is the generating process and the second part(0 to 50) is the inversion process. My question is, for the second graph, how did you get the initial latents of a real image at timestep 50 in the left side?

[bram-w]

Hello! Thanks for this question, it looks like this figure actually had a typo in it: the x-ticks on the bottom should be 0-50-0 instead of 50-0-50. Thanks for bringing this up!

…

On Wed, Nov 29, 2023 at 7:13 PM yanzengatsmu ***@***.***> wrote: I guess the graph in Figure 5 represents the changing of cosine similarity between x_t and y_t. From left to right, the first part(50 to 0) is the generating process and the second part(0 to 50) is the inversion process. My question is, for the second graph, how did you get the initial latents of a real image at timestep 50 in the left side? — Reply to this email directly, view it on GitHub <#14>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AXPAP2XE3EXSYLYS4YD57ZTYG72VXAVCNFSM6AAAAABAAPSAM6VHI2DSMVQWIX3LMV43ASLTON2WKOZSGAYTONZZGI4TQNQ> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[yanzengatsmu]

Thanks for your reply.