Open-Sora-Plan

Open-Sora-Plan copied to clipboard

Does this model support image-to-video inference?

4

support condition image as start frame

trouble-maker007

about sp seed config. Why seed in each device is different?

13

https://github.com/PKU-YuanGroup/Open-Sora-Plan/blob/main/opensora/npu_config.py#L336 why seed += self.rank? why not just use same seed in each device?

Edwardmark

Issues with Training CausalVideoVAE v1.2.0

5

Thank you for your wonderful work! I encountered several issues while trying to train the CausalVideoVAE model v1.2.0 using the provided training script `bash scripts/causalvae/train.sh.` The main problem is related...

Bravo5542

What does this PR do? --- This PR adds type checking to all functions for sampling a image or a video.

kimx3966

Sequence Parallelism Aware Training Loss

1

Hi there! When you are training with sequence parallel attention, I was wondering if you scale the loss function properly, as each GPU card will only contain a subset of...

ginward

Is DeepSpeed-Ulysses the sequence parallel method used in Open-Sora-Plan v1.2.0?

1

Hi, thank you for the great work! I was wondering is DeepSpeed-Ulysses the sequence parallel method used in both inference/training of Open-Sora-Plan v1.2.0? (As a side note, I think you...

TheTinyTeddy

What is the parameter count of the entire model？

3

What are the training parameters for videogpt and LatteT2V in the model？

atomrun39

Why not set random seed when running the inference?

3

Hi, thanks for your great work. I have found a strange thing, if I set the seed in the for loop, the video will be noise. And why you not...

Edwardmark

3D DIT pre-training weights？

2

Great work, I was wondering if the latest 3D DIT's can load pixart-alpha weights or other open source text2image model weights? Or it only can be trained from scratch?

RanXu2000

3dvae的训练loss是什么 ？最基本的vaeloss？encode和decode一起训练？

3

henbucuoshanghai