Inference time and memory use

[ZiyangYan]

Hi, thanks for your interesting work, I tried to do a inference with your script "sample_image.sh" . It took more than 1H to run the code and finally just got a OOM (I use 1 80G H100). Can you give me more information about the inference time and GPU uasge about running the inference?

[stdstu12]

Apologies for the delayed response; I've been occupied with my graduation and preparing for the v1.5 release.

In our tests using an A100 80G, we were able to run inference for 30-second videos on the 14B model without running out of memory (OOM). Could you please double-check your configuration? Alternatively, you might want to try using Yume-5B.