Description
I meet the error Error[3]: IExecutionContext::executeV2: Error Code 3: API Usage Error (Parameter check failed, condition: nullPtrAllowed. Tensor “images” is bound to nullptr, which is allowed only for an empty input tensor, shape tensor, or an output tensor associated with an IOuputAllocator.) while generate an int8 trtengine with custom calib file.
I used the same onnx and cache file and successfully converted the model on Xavier NX (Tensorrt 8.5.2 Jetpack 5.1.1). But when I copy those files to the Orin NX device (detail provided below), I meet this error. The onnx model is fine, it could be converted to trtengine if I didn’t specify the cache file.
Environment
TensorRT Version: 10.3.0.0
Device: Jetson Orin NX
Nvidia Driver Version: Jetpack 6.2 L4T 36.4.3
CUDA Version: 12.6
Steps To Reproduce
My command to generate trtengine:
trtexec --onnx=yolov7_trimmed_qat_noqdq.onnx --useDLACore=0 --int8 --inputIOFormats=int8:dla_hwc4 --outputIOFormats=int8:dla_linear --calib=/home/nvidia/workspace/chinnl/int8_quantize/yolov7/yolov7_trimmed_qat_precision_config_calib.cache
I uploaded the model files I used on GG drive: models - Google Drive
And the log file is attached.
log.txt (364.8 KB)
Please help! Thanks!