GazeSAM ONNX files + INT8 calibration caches
#6
by
nstlz
- opened
ONNX files for all model components. There are two YOLO files - yolo_m.onnx for TensorRT engine compilation, yolo_m_ort for running with ONNXRuntime. The difference between the two lies in the formatting of the NMS module.
The two TensorRT INT8 calibration caches provided are generated using the EntropyCalibrator2 algorithm.
han-cai
changed pull request status to
merged