Skip to content

Can not reproduce the results in paper #12

@fanfanffff1

Description

@fanfanffff1

I use VC2 following exact steps described in README, and bash script_sh/inference_t2v.sh to inference from prompts/vbench_standard_prompts.txt, but the generated video quality is not good. Does the code or config in the inference section need to be modified?

The generated results are as below:

VC2

https://github.com/user-attachments/assets/0e0ed18b-0c30-4256-9c9d-fd4c7315831c
https://github.com/user-attachments/assets/28302348-3cca-49cf-85a8-e44f0965ba2d

VC2+dpo

https://github.com/user-attachments/assets/60d2cda0-69d4-40e3-8677-c982d04f6ea7
https://github.com/user-attachments/assets/e3f4062e-5ded-4ffb-9436-a37755d461bf

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions