search

showlab / Tune-A-Video

[ICCV 2023] Tune-A-Video: One-Shot Tuning of Image Diffusion Models for Text-to-Video Generation
https://tuneavideo.github.io
Apache License 2.0
4.15k stars 376 forks source link

Question about the results #22

Closed liangbingzhao closed 1 year ago

liangbingzhao commented 1 year ago

I followed your README, modifying nothing of your code. Here is the final results, not that promising. What's more, I find that sample-100, sample-200 and sample-300 did not change much. Can u share how to improve it?

sample-300

zhangjiewu commented 1 year ago

please refer to https://github.com/showlab/Tune-A-Video/issues/15

liangbingzhao commented 1 year ago

Thank you, it seems like the xformers installation problem