generation of DAVIS-training-mp4

[Rose0506]

Hi, MatiasI'm going to do my research in your FastDVDNet. I try to use other datasets together with the DAVIS-training-mp4 for training. However, when i generate my mp4 files and used them together with DAVIS-training-mp4 for training, error occured 'Assert on "codec_id_ == codec_id" failed: File MP4/train_little/bear.mp4 is not the same codec as previous files Stacktrace', the reason must be different codec used in my mp4 files with DAVIS-training-mp4. So i want to konw the codec used in generating DAVIS-training-mp4, or could you give me the codes for preprocessing the DAVIS-training-mp4 in your github from jpg files. Sincerely hope to receive your reply! Thanks very much

[Rose0506]

Hi, MatiasI'm going to do my research in your FastDVDNet. I try to use other datasets together with the DAVIS-training-mp4 for training. However, when i generate my mp4 files and used them together with DAVIS-training-mp4 for training, error occured 'Assert on "codec_id_ == codec_id" failed: File MP4/train_little/bear.mp4 is not the same codec as previous files Stacktrace', the reason must be different codec used in my mp4 files with DAVIS-training-mp4. So i want to konw the codec used in generating DAVIS-training-mp4, or could you give me the codes for preprocessing the DAVIS-training-mp4 in your github from jpg files. Sincerely hope to receive your reply! Thanks very much

[m-tassano]

Hi @Rose0506 ,
You can have a look at https://gist.github.com/m-tassano/0536391eb79d63864e5005ea4da88243 and especially https://gist.github.com/m-tassano/2ec9ffa064484472de8ac7eaf66c2ed8 for the code used to generate the mp4.

Additionally, here you can find the list of Davis sequences in the trainset https://gist.github.com/m-tassano/27c1ef00ca42a8e50c2cee8a4205e559#file-davis_trainseq_list-txt

As for the codec used, I don't unfortunately have access to the machine on which I run the code, so I wouldn't be able to tell you specifically which codec was used.
Hope this helps.

[Rose0506]

Thanks very much! I also have another question about training details, how to divide the trainig-set and validation-set from the 90 mp4-files in The 2017 DAVIS? Thanks for your help again~

[m-tassano]

I used all the training videos in the list above for training only.
If you want, the Davis dataset has a separate validation set

[Rose0506]

By the way, could you recommend the code for generating the frame sequences from mp4? Just corresponding to the code you used to generate the DAVIS data mp4 from sequences.

[m-tassano]

I assume that you mean how to extract the frames from the mp4s. You can try something like
ffmpeg -skip_frame nokey -i input.mp4 -vsync vfr thumb%04d.png -hide_banner

(I haven't tried it but I believe it works)