PyTorch MobileNetV2SSD-Lite
·
30 commits
to develop
since this release
• Model has been optimized as a pre-processing for post quantization
• Batch norm folding and Adaround has been performed, this checkpoint does not contain batch norm layers
• This checkpoint has been evaluated in INT8 - mAP: 68.60%