Xingyi Yang

icon indicating a website link adamdad.github.io icon indicating an email [email protected]

icon company Hong Kong Polytechnic University @nus @torrvision @UCSD @seu icon location hong kong icon location Assistant Professor @ The Hong Kong Polytechnic University, Phd @ NUS xML Lab, Visiting Oxford, Machine learner

Results 90 comments of Xingyi Yang

how much memory needed？ oom error on 32G memory card

I use A6000 for my experiment. I am trying to figure out how to further save memory cost.🙏

Requirements file

I would be very grateful if you could share your version🙏. I haven't looked at the VideoCrafter repository in quite some time, so I might need a bit of time...

Conversion script (from Videocrafter to diffusers)

Thanks for the advice! That is actually an important technique components of our project (I guess one of the most important part). Will upload this week! Stay tuned!!!!!

Thank for your implementation of VideocrafterV2 in diffusers

Hello @shaoshitong . Thank you so much for your interest. I send a email to you.

support cogvideox

great question！will do shortly

Has anyone had this problem and what to do

Dear @mxy0610 , From your error code, it's unclear what settings you're using for your training. However, it seems you might be encountering issues due to using a single GPU...

RuntimeError: memory format option is only supported by strided tensors

Could you please provide more context about the problem? For example, the environment version, system details, and any other relevant information. Also, are you using `ROCm` devices?

The kat model training time problem

Can you see your GPU utilization? This is still slow. I use 8xA5000 and train around 1-2 days for 300 epoches.

RuntimeError: CUDA error: an illegal memory access was encountered Compile with `TORCH_USE_CUDA_DSA` to enable device-side assertions.

Hello, @1-2-3-4-0 @zxk72 someone also report this problem to me, and i'm looking for solutions. Can YOU please also look at this issue?https://github.com/pytorch/pytorch/issues/21819

RuntimeError: CUDA error: an illegal memory access was encountered Compile with `TORCH_USE_CUDA_DSA` to enable device-side assertions.

@bravoYJ Hello, my intuition is that the group number should not be set too large, as it may negate the parameter-saving benefits. I have not tested scenarios where the group...