tensorrt_inference icon indicating copy to clipboard operation
tensorrt_inference copied to clipboard

INT8 problem

Open PowerDi opened this issue 4 years ago • 0 comments

Could we perform ONNX model with INT8 mode? I defined a calibrator when I used layer API to build the model and it works well. But I use the same calibrator on onnx model it failed to output right result. Is it the same with building the model layer by layer? -yolov5

PowerDi avatar Jun 10 '21 11:06 PowerDi