PKU-YuanGroup / Video-LLaVA

【EMNLP 2024🔥】Video-LLaVA: Learning United Visual Representation by Alignment Before Projection
https://arxiv.org/pdf/2311.10122.pdf
Apache License 2.0
3.04k stars 220 forks source link

推理过程中出现错误 #117

Closed zhangye0402 closed 9 months ago

zhangye0402 commented 9 months ago
屏幕截图 2024-02-28 150612 屏幕截图 2024-02-28 150644

在推理过程中出现了如下报错,我修改了model_path到本地地址,同时修改了权重文件中的config.json中的mm_image_tower 以及 mm_video_tower 到本地地址,之后在推理中出现了图片上显示错误,请问我该如何修改才能正确运行video-llava?万分感谢!

zhangye0402 commented 9 months ago

我的,我翻到解决办法了

qingchunlizhi commented 8 months ago

我的,我翻到解决办法了

你是怎么解决的

liury889 commented 8 months ago

同问,我一样的问题