Generating TensorRT engine from TAO deploy


When converting models using tao deploy from onnx to tensorrt engine if we need to convert to FR16. Will it convert to FP32 and convert to FP16 or will it convert directly to FP16? If there are any documentations to get further information on conversion, it will be highly appreciated.

Thank you.

Convert to FP16 directly. You can refer to doc : YOLOv4 with TAO Deploy - NVIDIA Docs or

There is no update from you for a period, assuming this is not an issue anymore. Hence we are closing this topic. If need further support, please open a new one. Thanks