执行i2v模式时，出现错误RuntimeError: Sizes of tensors must match except in dimension 1. Expected size 16 but got size 8 for tensor number 1 in the list.

[hhue]

System Info / 系統信息

cuda12.2，python3.10，linux redhat7.5。相关依赖包及版本：transformer4.45.2，diffusers>=0.30.3，accelerate>=0.34.0，imageio-ffmpeg>=0.5.1

Information / 问题信息

• [X] The official example scripts / 官方的示例脚本
• [X] My own modified scripts / 我自己修改的脚本和任务

Reproduction / 复现过程

执行python cli_demo.py --prompt "A girl driver a car." --image_or_video_path "/data/llm/model/video/tmp/fpso.png" --generate_type "i2v" 出现如下错误 RuntimeError: Sizes of tensors must match except in dimension 1. Expected size 16 but got size 8 for tensor number 1 in the list. 如何解决呢，谢谢！

Expected behavior / 期待表现

期望尽早解决，我自己也在找办法，谢谢

[zRzRzRzRzRzRzR]

Did you download the T2V model instead of the I2V model? Please check CogVideoX-5B-I2V

[hhue]

已解决，找到原因了，需要用CogVideoX-5b-I2V模型，谢谢