taojishou

2022-09-18 23:22:49 | NEW | Input PID 0x0000 appeared (PAT) 2022-09-18 23:22:49 | *** | ERROR: Can't detect ECM pid. 2022-09-18 23:22:49 | NEW | Input PID 0x0610 appeared (PMT)...

./tsdecrypt --camd-proto NEWCAMD --camd-server 127.0.0.1:10100 --camd-user 1 --camd-pass 1 --camd-des-key 0102030405060708091011121314 --input 235.1.1.1:1234 --output 238.1.1.1:1234 --caid 0x0100 --ca-system SECA The same goes for the other CA encryption

 It seems to be ineffective, with the addition of int8 parameters, the model size has not changed much yolo export model=/home/tao/runs/detect/train9/weights/best.pt format=onnx int8=True

model.0.conv.weight 1 model.0.conv.bias 1 model.1.conv.weight 1 model.1.conv.bias 1 model.2.cv1.conv.weight 1 model.2.cv1.conv.bias 1 model.2.cv2.conv.weight 1 model.2.cv2.conv.bias 1 model.2.m.0.cv1.conv.weight 1 model.2.m.0.cv1.conv.bias 1 model.2.m.0.cv2.conv.weight 1 model.2.m.0.cv2.conv.bias 1 model.3.conv.weight 1 model.3.conv.bias 1 model.4.cv1.conv.weight 1...

when I output OpenVino format, it's ok

But I need Int8 quantization in onnx format~ not openvino format onnx Arguments imgsz, half, dynamic, simplify, opset, batch , no int8

Hope future support ONNX Int8 quantization with YOLOv8