For industrial deployment, it has been common practice to adopt quantization to further speed up runtime without much performance compromise. However, due to the heavy use of re-parameterization blocks in YOLOv6, previous PTQ techniques fail to produce high performance, while it is hard to incorporate QAT when it comes to matching fake quantizers during training and inference.
In order to solve the quantization problem of YOLOv6, we firstly reconstruct the network with RepOptimizer, and then perform well-designed PTQ and QAT skills on this model. Finally we can obtain a SOTA quantized result(mAP 43.3 at 869 QPS) for YOLOv6s.
Specific tutorials, please refer to the following links:
Dear OpenI User
Thank you for your continuous support to the Openl Qizhi Community AI Collaboration Platform. In order to protect your usage rights and ensure network security, we updated the Openl Qizhi Community AI Collaboration Platform Usage Agreement in January 2024. The updated agreement specifies that users are prohibited from using intranet penetration tools. After you click "Agree and continue", you can continue to use our services. Thank you for your cooperation and understanding.
For more agreement content, please refer to the《Openl Qizhi Community AI Collaboration Platform Usage Agreement》