Bug in Classify Videos Using Deep Learning

1 view (last 30 days)
TIENYU HUANG
TIENYU HUANG on 11 Jun 2019
Answered: Kenta on 31 Mar 2020
I download HMDB which has 51 action labled dataset as example requests. My hardware is Nvidia RTX Titan which has 24GB on graphic card. Platform is ubuntu 18.04 , Matlab 2019b, RAM 16GB on mainboard, intel i7 CPU. But I can't run through all action. Everytime, I stop at the line video = readVideo(files(i)); and the i is around 349 or 350. It hangs at inside readVideo and at vr = VideoReader(filename);
If I trained only two or three actions , which total video is not over 300+, the program runs well, and give the correct prediction.

Answers (1)

Kenta
Kenta on 31 Mar 2020
I think all of data could be down loaded propaery or something. I am not sure because the error statement is not attached.
The example requires comparatively big data and you have to down load it.
The demo below is also for video classification, but the dataset is light and easy to run.
It would be helpful to try that first. I hope it helps you a lot.

Categories

Find more on Startup and Shutdown in Help Center and File Exchange

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!