search

stanford-futuredata / noscope

Accelerating network inference over video
http://dawn.cs.stanford.edu/2017/06/22/noscope/
436 stars 122 forks source link

Use my own video to run NoScope #36

Closed janhuang6 closed 6 years ago

janhuang6 commented 6 years ago

First my video TimeSquare.mp4 is 29 minutes, is that long enough for NoScope study? How to create the TimeSquare.csv required by NoScope?

Thanks!

ddkang commented 6 years ago

All of our experiments used at least 8 hours of training video, but you are welcome to experiment with less.

You can create the necessary training data using yolo_standalone

janhuang6 commented 6 years ago

Do you have the URL for yolo_standalone? or where to get it?

Thanks!

sxhexe commented 6 years ago

I think you can get it from compiling darknet at tensorflow-noscope/tensorflow/noscope/darknet.

janhuang6 commented 6 years ago

Thank you for the input. This issue can be closed now.