search

triple-Mu / YOLOv8-TensorRT

YOLOv8 using TensorRT accelerate !
MIT License
1.32k stars 228 forks source link

Input size variety. #200

Open ablyzniuk opened 6 months ago

ablyzniuk commented 6 months ago

How to keep different picture sizes(I meant dynamic input shape size, not only alongside batch axis) as the original ultralytics approach describes?

triple-Mu commented 6 months ago

How to keep different picture sizes(I meant dynamic input shape size, not only alongside batch axis) as the original ultralytics approach describes?

If you want dynamic shape input size, you'd better export an onnx with dynamic shape. But there will be some difference in inference code. Such as in cudaMalloc and set trt bindings.