search

facebookarchive / C3D

C3D is a modified version of BVLC caffe to support 3D ConvNets.
Other
1.16k stars 507 forks source link

How to output the accuracy of a single input video? #385

Closed Jessespace closed 6 years ago

Jessespace commented 6 years ago

How to output the accuracy of a single input video?

Jessespace commented 6 years ago

like this

image

dutran commented 6 years ago

This one is not supported in the current repo, you can use feature_extraction tool extract softmax prediction for clips (densely from a video).