zero output

[DidiD1]

I want to use this to evaluate my own videos, but it's strange that the output is all 0. Could you help me to correctly use the metricks to eval, thanks!

[DidiD1]

the code i use:

python evaluate.py \
 --videos_path ./VBench/11_29_3s \
 --dimension "motion_smoothness" \
 --mode "custom_input"

and the output be like:
{
"subject_consistency": [
0.0,
[
{
"video_path": xxx,
"video_results": 0.0
},
{
"video_path":xxx,
"video_results": 0.0
},
{
"video_path": xxx,
"video_results": 0.0
},

[rahimentezari]

same problem here #87

[DidiD1]

same problem here #87
I track the bug found a strange thing is that, the image_features of the clip is all 0, which leads to the result of compute_background_consistency be zero. I don't know what's the problem, maybe the version of package?

[NattapolChan]

Can you provide ls -lh ~/.cache/vbench/* and version here? Seems like the problem is not specific to CLIP as subject_consistency does not use it (#87).

[zhongyu-zhao]

same problem here #87
I track the bug found a strange thing is that, the image_features of the clip is all 0, which leads to the result of compute_background_consistency be zero. I don't know what's the problem, maybe the version of package?

I encountered the same problem and found the same thing as you. Besides, I noticed that the first model inference process took a unusually long time(approximately 10-12minutes) while other steps could finish in just minutes. Have you solved this bug?