使用从头训练的模型生成的视频画面一致性较差

[kawayi12318]

sample_0.mp4

使用数据集：HD-VG-130M
step：136500
batch_size：1

请问这是什么原因？

[lanson07]

Hi~
Which model did you use for inference？

[github-actions]

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

[zhengzangw]

你用单卡训练，global batch size 是 1 吗？这样的话过的数据还是太少了，而且 batch size 太小，最后模型的效果也可能不够理想。

[github-actions]

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

[github-actions]

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