MixFormerV2: Efficient Fully Transformer Tracking.

official pytorch

Here, the c++ version of onnx mixformerv2 tracking algorithm is provided, and the fps reaches about 300fps on the 3080-laptop gpu.

At the same time, a pytrt version was also provided, which reached 430fps on the 3080-laptop gpu.

mixformer_v2.onnx

mixformer_v2_sim.onnx

mixformer_v2_sim.engine

Prerequisites: First, download the source code of onnx and compile it. For details, see lite.ai.toolkit. Put the header file into the onnxruntime folder and put the compiled .so file into the lib folder. The above two folders are located in Mixformerv2-onnx. However, the above steps are not required for tensorRT inference, you only need to configure TensorRT.

modify onnx path as yours

$ mkdir build && cd build
$ cmake .. && make -j

$ cd /home/code/Mixformerv2-onnx
$ ./mixformer-onnx [model_path] [videopath(file or camera)]

Modify the video path in Mixformerv2-onnx/mixformer-pytrt/mf_tracker_trt.py，and mkdir model file_dir, then download the onnx file and put onnx file into file_dir.

$ cd Mixformerv2-onnx
& python mixformer-pytrt/onnx2trt.py 
$ python mixformer-pytrt/mf_tracker_trt.py

Note: In addition to simplification when converting the onnx model, it is important to ensure that the shape of the data input to the engine model and the corresponding underlying data are continuous.

Thanks for the LightTrack-ncnn and lite.ai.tookit, which helps us to quickly implement our ideas.