search

hpcaitech / Open-Sora

Open-Sora: Democratizing Efficient Video Production for All
https://hpcaitech.github.io/Open-Sora/
Apache License 2.0
21.52k stars 2.06k forks source link

the dimensions of input video? #622

Closed henbucuoshanghai closed 1 month ago

henbucuoshanghai commented 1 month ago

https://github.com/hpcaitech/Open-Sora/blob/476b6dc79720e5d9ddfb3cd589680b2308871926/opensora/models/stdit/stdit3.py#L364C8-L364C43 T, H, W = self.get_dynamic_size(x)

    S = H * W
    base_size = round(S**0.5)
    resolution_sq = (height[0].item() * width[0].item()) ** 0.5
    scale = resolution_sq / self.input_sq_size
    pos_emb = self.pos_embed(x, H, W, scale=scale, base_size=base_size)
henbucuoshanghai commented 1 month ago

T, H, W = self.get_dynamic_size(x) even in one batch, the input video embedding of T H W is different with each other? every video has its own T W H? so the input dimensions of STdit3.py is changee depend on every video?

henbucuoshanghai commented 1 month ago

视频vae压缩后，维度一致？

github-actions[bot] commented 1 month ago

This issue is stale because it has been open for 7 days with no activity.

github-actions[bot] commented 1 month ago

This issue was closed because it has been inactive for 7 days since being marked as stale.