Matt Hill
Matt Hill
Comparing mobilenet v1 and v2 for **inferencing** on the **CPU**, I have observed some surprising numbers: 1. For v1, my inference time was about 148ms on average. For **v2, the...
This change https://github.com/intel/caffe/blame/7010334f159da247db3fe3a9d96a3116ca06b09a/src/caffe/util/bbox_util.cpp#L2266 using `at` instead of `push_back` is selectively over-writing some results, and leaving some low-scoring results within the top-k entries if there are not many strong detections in...