search

MichiganCOG / ViP

Video Platform for Action Recognition and Object Detection in Pytorch
MIT License
219 stars 37 forks source link

Update Extract Clips #36

Closed ehofesmann closed 4 years ago

ehofesmann commented 4 years ago

Added numclips>0, clip stride, and clip_length=-1 options to extract clips. Also updated all current datasets to allow for variable length video loading. Parse args was changed to set any clip_stride that is < 1 to 1.

ehofesmann commented 4 years ago

Addressing issues #29 and #26

natlouis commented 4 years ago

Added assertion for batch_size = 1 and clip_length = -1