When I load onnxfile by deepstream,
error message is :
gst-launch-1.0 rtspsrc location=rtsp://admin:admin@192.168.3.100:8557/h264 ! rtph264depay ! nvv4l2decoder ! m.sink_0 nvstreammux name=m batch-size=1 width=1920 height=1080 ! nvinfer config-file-path=tgie_config.txt ! nvvideoconvert ! nvdsosd ! nvvideoconvert ! nvv4l2h264enc ! h264parse ! flvmux ! rtmpsink location=rtmp://192.168.1.213:1935/live/2020 sync=false
Setting pipeline to PAUSED …
Opening in BLOCKING MODE
Opening in BLOCKING MODE
0:00:00.422628374 14487 0x55c353af20 INFO nvinfer gstnvinfer.cpp:619:gst_nvinfer_logger: NvDsInferContext[UID 10003]: Info from NvDsInferContextImpl::buildModel() <nvdsinfer_context_impl.cpp:1719> [UID = 10003]: Trying to create engine from model files
Input filename: /home/nvidia/luozw/tensorRT-7/data/onnx/smoke_phone.onnx
ONNX IR version: 0.0.6
Opset version: 11
Producer name: tf2onnx
Producer version: 1.5.5
Domain:
Model version: 0
Doc string:
WARNING: [TRT]: onnx2trt_utils.cpp:220: Your ONNX model has been generated with INT64 weights, while TensorRT does not natively support INT64. Attempting to cast down to INT32.
WARNING: INT8 calibration file not specified. Trying FP16 mode.
WARNING: [TRT]: DLA requests all profiles have same min, max, and opt value. All dla layers are falling back to GPU
INFO: [TRT]:
INFO: [TRT]: --------------- Layers running on DLA:
INFO: [TRT]:
INFO: [TRT]: --------------- Layers running on GPU:
INFO: [TRT]: darknet/conv0/Conv2D__20, Conv__247, darknet/conv0/LeakyRelu, Conv__248, darknet/conv1/LeakyRelu, Conv__251, darknet/residual0/conv1/LeakyRelu, Conv__252, PWN(darknet/residual0/conv2/LeakyRelu, darknet/residual0/add), Conv__253, darknet/conv4/LeakyRelu, Conv__256, darknet/residual1/conv1/LeakyRelu, Conv__257, PWN(darknet/residual1/conv2/LeakyRelu, darknet/residual1/add), Conv__260, darknet/residual2/conv1/LeakyRelu, Conv__261, PWN(darknet/residual2/conv2/LeakyRelu, darknet/residual2/add), Conv__262, darknet/conv9/LeakyRelu, Conv__265, darknet/residual3/conv1/LeakyRelu, Conv__266, PWN(darknet/residual3/conv2/LeakyRelu, darknet/residual3/add), Conv__269, darknet/residual4/conv1/LeakyRelu, Conv__270, PWN(darknet/residual4/conv2/LeakyRelu, darknet/residual4/add), Conv__273, darknet/residual5/conv1/LeakyRelu, Conv__274, PWN(darknet/residual5/conv2/LeakyRelu, darknet/residual5/add), Conv__277, darknet/residual6/conv1/LeakyRelu, Conv__278, PWN(darknet/residual6/conv2/LeakyRelu, darknet/residual6/add), Conv__281, darknet/residual7/conv1/LeakyRelu, Conv__282, PWN(darknet/residual7/conv2/LeakyRelu, darknet/residual7/add), Conv__285, darknet/residual8/conv1/LeakyRelu, Conv__286, PWN(darknet/residual8/conv2/LeakyRelu, darknet/residual8/add), Conv__289, darknet/residual9/conv1/LeakyRelu, Conv__290, PWN(darknet/residual9/conv2/LeakyRelu, darknet/residual9/add), Conv__293, darknet/residual10/conv1/LeakyRelu, Conv__294, PWN(darknet/residual10/conv2/LeakyRelu, darknet/residual10/add), Conv__297, darknet/conv26/LeakyRelu, Conv__300, darknet/residual11/conv1/LeakyRelu, Conv__301, PWN(darknet/residual11/conv2/LeakyRelu, darknet/residual11/add), Conv__304, darknet/residual12/conv1/LeakyRelu, Conv__305, PWN(darknet/residual12/conv2/LeakyRelu, darknet/residual12/add), Conv__308, darknet/residual13/conv1/LeakyRelu, Conv__309, PWN(darknet/residual13/conv2/LeakyRelu, darknet/residual13/add), Conv__312, darknet/residual14/conv1/LeakyRelu, Conv__313, PWN(darknet/residual14/conv2/LeakyRelu, darknet/residual14/add), Conv__316, darknet/residual15/conv1/LeakyRelu, Conv__317, PWN(darknet/residual15/conv2/LeakyRelu, darknet/residual15/add), Conv__320, darknet/residual16/conv1/LeakyRelu, Conv__321, PWN(darknet/residual16/conv2/LeakyRelu, darknet/residual16/add), Conv__324, darknet/residual17/conv1/LeakyRelu, Conv__325, PWN(darknet/residual17/conv2/LeakyRelu, darknet/residual17/add), Conv__328, darknet/residual18/conv1/LeakyRelu, Conv__329, PWN(darknet/residual18/conv2/LeakyRelu, darknet/residual18/add), darknet/conv43/Conv2D, darknet/conv43/LeakyRelu, Conv__334, darknet/residual19/conv1/LeakyRelu, darknet/residual19/conv2/Conv2D, PWN(darknet/residual19/conv2/LeakyRelu, darknet/residual19/add), Conv__337, darknet/residual20/conv1/LeakyRelu, darknet/residual20/conv2/Conv2D, PWN(darknet/residual20/conv2/LeakyRelu, darknet/residual20/add), Conv__340, darknet/residual21/conv1/LeakyRelu, darknet/residual21/conv2/Conv2D, PWN(darknet/residual21/conv2/LeakyRelu, darknet/residual21/add), Conv__343, darknet/residual22/conv1/LeakyRelu, darknet/residual22/conv2/Conv2D, PWN(darknet/residual22/conv2/LeakyRelu, darknet/residual22/add), Conv__344, conv52/LeakyRelu, conv53/Conv2D, conv53/LeakyRelu, Conv__345, conv54/LeakyRelu, conv55/Conv2D, conv55/LeakyRelu, Conv__346, conv56/LeakyRelu, Conv__350, conv_lobj_branch/Conv2D, conv57/LeakyRelu, conv_lobj_branch/LeakyRelu, Conv__349, conv_lbbox/Conv2D__143 + pred_lbbox/reshape, pred_lbbox/strided_slice, pred_lbbox/Sigmoid, pred_lbbox/Reshape_1, pred_lbbox/strided_slice_1, pred_lbbox/Exp, pred_lbbox/Reshape_2, pred_lbbox/strided_slice_2, pred_lbbox/Sigmoid_1, pred_lbbox/strided_slice_3, pred_lbbox/Sigmoid_2, Resize__172, Resize__172:0 copy, Conv__351, conv58/LeakyRelu, Conv__352, (Unnamed Layer* 166) [Constant] + pred_lbbox/Add + (Unnamed Layer* 168) [Constant] + pred_lbbox/Mul, pred_lbbox/Reshape_4, (Unnamed Layer* 158) [Constant] + pred_lbbox/Mul_1 + (Unnamed Layer* 160) [Constant] + pred_lbbox/Mul_2, pred_lbbox/Reshape_5, pred_lbbox/Reshape_4:0 copy, pred_lbbox/Reshape_5:0 copy, pred_lbbox/concat:0 copy, pred_lbbox/S
INFO: [TRT]: Some tactics do not have sufficient workspace memory to run. Increasing workspace size may increase performance, please check verbose output.
ERROR: [TRT]: …/builder/cudnnBuilderUtils.cpp (427) - Cuda Error in findFastestTactic: 700 (an illegal memory access was encountered)
ERROR: [TRT]: Parameter check failed at: /home/jenkins/workspace/TensorRT/helpers/rel-7.1/L1_Nightly_Internal/build/source/rtSafe/resources.h::operator()::393, condition: CudaDeleterAPI(ptr) failure.
ERROR: [TRT]: …/rtSafe/safeRuntime.cpp (32) - Cuda Error in free: 700 (an illegal memory access was encountered)
terminate called after throwing an instance of ‘nvinfer1::CudaError’
what(): std::exception
Aborted (core dumped)
When I load onnxfile by tensorrt ;it works well.
How can I solved this problem
• Hardware Platform (Jetson / GPU)
• DeepStream Version5.0
• JetPack Version (valid for Jetson only)
• TensorRT Version
• NVIDIA GPU Driver Version (valid for GPU only)