NVIDIA Developer Forums

Is it possible to feed unsigned char data into TensorRT rather than float32?

AI & Data Science Deep Learning (Training & Inference) TensorRT

SunilJB March 23, 2020, 5:01am 2

Hi,

TensorRT only supports FP32, FP16, INT32, and INT8: DataType — NVIDIA TensorRT Standard Python API Documentation 8.6.1 documentation

Please refer to this similar topic:

Thanks

Topic		Replies	Views	Activity
Input images with dtype int8 TensorRT	5	1586	January 17, 2021
How to transfer int8 to float32, and do not use int8 calibrate TensorRT	3	998	August 11, 2020
How to pass uint8 input to a tensorrt engine? TensorRT	6	2873	October 12, 2021
How can I use int8 typed inputs into TensorRT6.0? TensorRT	4	1064	June 19, 2020
TensorRT INT8 calibration in C++ api TensorRT tensorrt	2	1821	February 14, 2022
Giving input in NHWC format to tensor rt network in trt7 TensorRT	1	666	September 30, 2020
Can TensorRT 7.1.3 convert an INT8 pytorch QAT model to engine? TensorRT	3	745	April 21, 2022
tensorrt inference error while load onnx model TensorRT	8	3371	October 12, 2021
TensorRT quantization uses int8 or uint8 TensorRT tensorrt	1	859	June 6, 2023
Data inferencing to INT8U quantized model TensorRT tensorrt	2	411	October 12, 2021