Data doesn't store properly
Hi,
what is the shape of the data in npy file of every video?
I tried but it doesn't store in a proper format.
Can you guide me about it?

Hi, As you can see from the log you have decoding errors: ffprobe failed at: ... Which means there is something wrong with your videos (either the videos are corrupted or you provided the wrong paths)
Thanks, man. solved by giving relative path.
but the data is stored in npy file which looks like this. I am sure this not what you got right.
How can I give this data as input into model(CNN +distributed layer for video classification)

That's weird, you should get arrays of shape Nx1024 with N > 0. Did you make sure this was not the numpy arrays that were output from the first run which had a problem? If yes you might need to delete these generated npy arrays and rerun the extraction with the path fixed.
I did but doesn't store anything in npy file. I am sure it doesn't look like any logical error but it might be the problem of video files. I don't know about it.
There must be something wrong with the video decoding. Did you try playing the videos? Or better dumping the frames with ffmpeg?
@pythonmjs can you please show me how did you build you csv file ? i am getting issue with it
Hey @antoine77340 , nothing is saved in bumpy files in my case. Can you please help with this?
Hello,
you can refer to my project, but there are some garbage files in it. I didn't clean them up when I did it again you can do follow it python input_ file. py python extract. py
------------------ 原始邮件 ------------------ 发件人: "antoine77340/video_feature_extractor" @.>; 发送时间: 2022年4月12日(星期二) 下午3:30 @.>; @.@.>; 主题: Re: [antoine77340/video_feature_extractor] Data doesn't store properly (#8)
dear @antoine77340 ,
i tried running sample of video segments on my EC2 ubuntu, output as below;
however, result in my video_feature folder is empty. Nothing is there. My csv is as below;
Do all the videos and feature folder need to be inside the github folder to make this work? Or can you advice on this issue. Much appreciated!
— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you commented.Message ID: @.***>