[JETSON XAVIER] ERROR from primary_gie_classifier: NvBufSurfTransform failed with error -2 while con...

Hi.

I need help to fix this ERROR from primary_gie_classifier: NvBufSurfTransform failed with error -2 while converting buffer error.

I’m running deepstream-app -c source12_1080p_dec_infer-resnet_tracker_tiled_display_fp16_tx2_teste.txt

Creating LL OSD context new
0:00:06.669589624  4288      0x546ee70 WARN                 nvinfer gstnvinfer.cpp:515:gst_nvinfer_logger:<primary_gie_classifier> NvDsInferContext[UID 1]:checkEngineParams(): Requested Max Batch Size is less than engine batch size
0:00:06.673240148  4288      0x546ee70 INFO                 nvinfer gstnvinfer.cpp:519:gst_nvinfer_logger:<primary_gie_classifier> NvDsInferContext[UID 1]:initialize(): Trying to create engine from model files
0:00:06.674043382  4288      0x546ee70 WARN                 nvinfer gstnvinfer.cpp:515:gst_nvinfer_logger:<primary_gie_classifier> NvDsInferContext[UID 1]:generateTRTModel(): INT8 calibration file not specified. Trying FP16 mode.
0:00:15.643457716  4288      0x546ee70 INFO                 nvinfer gstnvinfer.cpp:519:gst_nvinfer_logger:<primary_gie_classifier> NvDsInferContext[UID 1]:generateTRTModel(): Storing the serialized cuda engine to file at /usr/src/tensorrt/samples/python/end_to_end_tensorflow_mnist/models/lenet5.uff_b8_fp16.engine

Runtime commands:
	h: Print this help
	q: Quit

	p: Pause
	r: Resume


**PERF: FPS 0 (Avg)	FPS 1 (Avg)	FPS 2 (Avg)	FPS 3 (Avg)	FPS 4 (Avg)	FPS 5 (Avg)	FPS 6 (Avg)	FPS 7 (Avg)	FPS 8 (Avg)	FPS 9 (Avg)	FPS 10 (Avg)	FPS 11 (Avg)	
**PERF: 0.00 (0.00)	0.00 (0.00)	0.00 (0.00)	0.00 (0.00)	0.00 (0.00)	0.00 (0.00)	0.00 (0.00)	0.00 (0.00)	0.00 (0.00)	0.00 (0.00)	0.00 (0.00)	0.00 (0.00)	
** INFO: <bus_callback:189>: Pipeline ready

Opening in BLOCKING MODE 
Opening in BLOCKING MODE 
Opening in BLOCKING MODE 
Opening in BLOCKING MODE 
Opening in BLOCKING MODE 
Opening in BLOCKING MODE 
NvMMLiteOpen : Block : BlockType = 261 
NvMMLiteOpen : Block : BlockType = 261 
NvMMLiteOpen : Block : BlockType = 261 
Opening in BLOCKING MODE 
NvMMLiteOpen : Block : BlockType = 261 
NvMMLiteOpen : Block : BlockType = 261 
Opening in BLOCKING MODE 
Opening in BLOCKING MODE 
NvMMLiteOpen : Block : BlockType = 261 
NvMMLiteOpen : Block : BlockType = 261 
NVMEDIA: Reading vendor.tegra.display-size : status: 6 
NvMMLiteOpen : Block : BlockType = 261 
Opening in BLOCKING MODE 
Opening in BLOCKING MODE 
NVMEDIA: Reading vendor.tegra.display-size : status: 6 
NVMEDIA: Reading vendor.tegra.display-size : status: 6 
NVMEDIA: Reading vendor.tegra.display-size : status: 6 
NVMEDIA: Reading vendor.tegra.display-size : status: 6 
NvMMLiteOpen : Block : BlockType = 261 
NVMEDIA: Reading vendor.tegra.display-size : status: 6 
Opening in BLOCKING MODE 
NvMMLiteOpen : Block : BlockType = 261 
NvMMLiteOpen : Block : BlockType = 261 
NvMMLiteOpen : Block : BlockType = 261 
NvMMLiteBlockCreate : Block : BlockType = 261 
NVMEDIA: Reading vendor.tegra.display-size : status: 6 
NVMEDIA: Reading vendor.tegra.display-size : status: 6 
NvMMLiteBlockCreate : Block : BlockType = 261 
NvMMLiteBlockCreate : Block : BlockType = 261 
NvMMLiteBlockCreate : Block : BlockType = 261 
NvMMLiteBlockCreate : Block : BlockType = 261 
NvMMLiteBlockCreate : Block : BlockType = 261 
NVMEDIA: Reading vendor.tegra.display-size : status: 6 
NvMMLiteBlockCreate : Block : BlockType = 261 
NVMEDIA: Reading vendor.tegra.display-size : status: 6 
NVMEDIA: Reading vendor.tegra.display-size : status: 6 
NVMEDIA: Reading vendor.tegra.display-size : status: 6 
NvMMLiteBlockCreate : Block : BlockType = 261 
NvMMLiteBlockCreate : Block : BlockType = 261 
NvMMLiteBlockCreate : Block : BlockType = 261 
NvMMLiteBlockCreate : Block : BlockType = 261 
NvMMLiteBlockCreate : Block : BlockType = 261 
Creating LL OSD context new
** INFO: <bus_callback:175>: Pipeline running

0:00:16.730014705  4288      0x4e4c720 WARN                 nvinfer gstnvinfer.cpp:1149:convert_batch_and_push_to_input_thread:<primary_gie_classifier> error: NvBufSurfTransform failed with error -2 while converting buffer
ERROR from primary_gie_classifier: NvBufSurfTransform failed with error -2 while converting buffer
Debug info: /dvs/git/dirty/git-master_linux/deepstream/sdk/src/gst-plugins/gst-nvinfer/gstnvinfer.cpp(1149): convert_batch_and_push_to_input_thread (): /GstPipeline:pipeline/GstBin:primary_gie_bin/GstNvInfer:primary_gie_classifier
Quitting
ERROR from qtdemux3: Internal data stream error.
Debug info: qtdemux.c(6073): gst_qtdemux_loop (): /GstPipeline:pipeline/GstBin:multi_src_bin/GstBin:src_sub_bin0/GstURIDecodeBin:src_elem/GstDecodeBin:decodebin0/GstQTDemux:qtdemux3:
streaming stopped, reason error (-5)
ERROR from qtdemux9: Internal data stream error.
Debug info: qtdemux.c(6073): gst_qtdemux_loop (): /GstPipeline:pipeline/GstBin:multi_src_bin/GstBin:src_sub_bin9/GstURIDecodeBin:src_elem/GstDecodeBin:decodebin9/GstQTDemux:qtdemux9:
streaming stopped, reason error (-5)
ERROR from qtdemux0: Internal data stream error.
Debug info: qtdemux.c(6073): gst_qtdemux_loop (): /GstPipeline:pipeline/GstBin:multi_src_bin/GstBin:src_sub_bin1/GstURIDecodeBin:src_elem/GstDecodeBin:decodebin1/GstQTDemux:qtdemux0:
streaming stopped, reason error (-5)
ERROR from qtdemux2: Internal data stream error.
Debug info: qtdemux.c(6073): gst_qtdemux_loop (): /GstPipeline:pipeline/GstBin:multi_src_bin/GstBin:src_sub_bin4/GstURIDecodeBin:src_elem/GstDecodeBin:decodebin4/GstQTDemux:qtdemux2:
streaming stopped, reason error (-5)
ERROR from qtdemux5: Internal data stream error.
Debug info: qtdemux.c(6073): gst_qtdemux_loop (): /GstPipeline:pipeline/GstBin:multi_src_bin/GstBin:src_sub_bin5/GstURIDecodeBin:src_elem/GstDecodeBin:decodebin5/GstQTDemux:qtdemux5:
streaming stopped, reason error (-5)
ERROR from qtdemux8: Internal data stream error.
Debug info: qtdemux.c(6073): gst_qtdemux_loop (): /GstPipeline:pipeline/GstBin:multi_src_bin/GstBin:src_sub_bin8/GstURIDecodeBin:src_elem/GstDecodeBin:decodebin8/GstQTDemux:qtdemux8:
streaming stopped, reason error (-5)
ERROR from qtdemux6: Internal data stream error.
Debug info: qtdemux.c(6073): gst_qtdemux_loop (): /GstPipeline:pipeline/GstBin:multi_src_bin/GstBin:src_sub_bin3/GstURIDecodeBin:src_elem/GstDecodeBin:decodebin3/GstQTDemux:qtdemux6:
streaming stopped, reason error (-5)
ERROR from multiqueue5: Internal data stream error.
Debug info: gstmultiqueue.c(2357): gst_multi_queue_sink_event (): /GstPipeline:pipeline/GstBin:multi_src_bin/GstBin:src_sub_bin3/GstURIDecodeBin:src_elem/GstDecodeBin:decodebin3/GstMultiQueue:multiqueue5:
streaming stopped, reason error (-5)
App run failed

the source file source12_1080p_dec_infer-resnet_tracker_tiled_display_fp16_tx2_teste.txt is a copy of the file source12_1080p_dec_infer-resnet_tracker_tiled_display_fp16_tx2.txt

source12_1080p_dec_infer-resnet_tracker_tiled_display_fp16_tx2_teste.txt

# Copyright (c) 2019 NVIDIA Corporation.  All rights reserved.
#
# NVIDIA Corporation and its licensors retain all intellectual property
# and proprietary rights in and to this software, related documentation
# and any modifications thereto.  Any use, reproduction, disclosure or
# distribution of this software and related documentation without an express
# license agreement from NVIDIA Corporation is strictly prohibited.

[application]
enable-perf-measurement=1
perf-measurement-interval-sec=5
#gie-kitti-output-dir=streamscl

[tiled-display]
enable=0
rows=4
columns=3
width=1280
height=720
gpu-id=0
#(0): nvbuf-mem-default - Default memory allocated, specific to particular platform
#(1): nvbuf-mem-cuda-pinned - Allocate Pinned/Host cuda memory, applicable for Tesla
#(2): nvbuf-mem-cuda-device - Allocate Device cuda memory, applicable for Tesla
#(3): nvbuf-mem-cuda-unified - Allocate Unified cuda memory, applicable for Tesla
#(4): nvbuf-mem-surface-array - Allocate Surface Array memory, applicable for Jetson
nvbuf-memory-type=0

[source0]
enable=1
#Type - 1=CameraV4L2 2=URI 3=MultiURI 4=RTSP
type=3
uri=file://../../streams/sample_1080p_h264.mp4
num-sources=12
#drop-frame-interval=2
gpu-id=0
# (0): memtype_device   - Memory type Device
# (1): memtype_pinned   - Memory type Host Pinned
# (2): memtype_unified  - Memory type Unified
cudadec-memtype=0

[sink0]
enable=1
#Type - 1=FakeSink 2=EglSink 3=File
type=5
sync=1
source-id=0
gpu-id=0
qos=0
nvbuf-memory-type=0
overlay-id=1

[sink1]
enable=0
type=3
#1=mp4 2=mkv
container=1
#1=h264 2=h265
codec=1
sync=0
#iframeinterval=10
bitrate=2000000
output-file=out.mp4
source-id=0

[sink2]
enable=0
#Type - 1=FakeSink 2=EglSink 3=File 4=RTSPStreaming
type=4
#1=h264 2=h265
codec=1
sync=0
bitrate=4000000
# set below properties in case of RTSPStreaming
rtsp-port=8554
udp-port=5400

[osd]
enable=1
gpu-id=0
border-width=1
text-size=15
text-color=1;1;1;1;
text-bg-color=0.3;0.3;0.3;1
font=Serif
show-clock=0
clock-x-offset=800
clock-y-offset=820
clock-text-size=12
clock-color=1;0;0;0
nvbuf-memory-type=0

[streammux]
gpu-id=0
##Boolean property to inform muxer that sources are live
live-source=0
batch-size=12
##time out in usec, to wait after the first buffer is available
##to push the batch even if the complete batch is not formed
batched-push-timeout=40000
## Set muxer output width and height
width=1200
height=1200
##Enable to maintain aspect ratio wrt source, and allow black borders, works
##along with width, height properties
enable-padding=0
nvbuf-memory-type=0


# config-file property is mandatory for any gie section.
# Other properties are optional and if set will override the properties set in
# the infer config file.
[primary-gie]
##enable=1
#gpu-id=0
#model-engine-file=/usr/src/tensorrt/bin/lenet5.engine
#batch-size=12
#Required by the app for OSD, not a plugin property
#bbox-border-color0=1;0;0;1
#bbox-border-color1=0;1;1;1
#bbox-border-color2=0;0;1;1
#bbox-border-color3=0;1;0;1
#interval=4
#gie-unique-id=1
#nvbuf-memory-type=0
enable=1
gpu-id=0
bbox-border-color0=1;0;0;1
bbox-border-color1=0;1;1;1
bbox-border-color2=0;0;1;1
bbox-border-color3=0;1;0;1
interval=4
gie-unique-id=1
nvbuf-memory-type=0
config-file=config_infer_primary_nano_teste.txt

[tracker]
enable=0
tracker-width=480
tracker-height=272
#ll-lib-file=/opt/nvidia/deepstream/deepstream-4.0/lib/libnvds_mot_iou.so
ll-lib-file=/opt/nvidia/deepstream/deepstream-4.0/lib/libnvds_mot_klt.so
#ll-config-file required for IOU only
#ll-config-file=iou_config.txt
gpu-id=0

[tests]
file-loop=0

the file config_infer_primary_nano_teste.txt is copy file the config_infer_primary_nano.txt. In my file model-color-format=2 when changed the parameter model-color-format = 2 to 0 or 1 the error changes, but I also can’t solve the problem

# Copyright (c) 2018 NVIDIA Corporation.  All rights reserved.
#
# NVIDIA Corporation and its licensors retain all intellectual property
# and proprietary rights in and to this software, related documentation
# and any modifications thereto.  Any use, reproduction, disclosure or
# distribution of this software and related documentation without an express
# license agreement from NVIDIA Corporation is strictly prohibited.

# Following properties are mandatory when engine files are not specified:
#   int8-calib-file(Only in INT8)
#   Caffemodel mandatory properties: model-file, proto-file, output-blob-names
#   UFF: uff-file, input-dims, uff-input-blob-name, output-blob-names
#   ONNX: onnx-file
#
# Mandatory properties for detectors:
#   num-detected-classes
#
# Optional properties for detectors:
#   enable-dbscan(Default=false), interval(Primary mode only, Default=0)
#   custom-lib-path,
#   parse-bbox-func-name
#
# Mandatory properties for classifiers:
#   classifier-threshold, is-classifier
#
# Optional properties for classifiers:
#   classifier-async-mode(Secondary mode only, Default=false)
#
# Optional properties in secondary mode:
#   operate-on-gie-id(Default=0), operate-on-class-ids(Defaults to all classes),
#   input-object-min-width, input-object-min-height, input-object-max-width,
#   input-object-max-height
#
# Following properties are always recommended:
#   batch-size(Default=1)
#
# Other optional properties:
#   net-scale-factor(Default=1), network-mode(Default=0 i.e FP32),
#   model-color-format(Default=0 i.e. RGB) model-engine-file, labelfile-path,
#   mean-file, gie-unique-id(Default=0), offsets, gie-mode (Default=1 i.e. primary),
#   custom-lib-path, network-mode(Default=0 i.e FP32)
#
# The values in the config file are overridden by values set through GObject
# properties.

    [property]
    gpu-id=0
    net-scale-factor=0.0039215697906911373
    batch-size=8
    process-mode=1
    model-color-format=2
    ## 0=FP32, 1=INT8, 2=FP16 mode
    network-mode=1
    num-detected-classes=10
    interval=0
    gie-unique-id=1
    #uff-file=/usr/src/tensorrt/samples/python/uff_custom_plugin/models/trained_lenet5.uff
    #uff-file=../../models/lenet5.uff
    uff-file=/usr/src/tensorrt/samples/python/end_to_end_tensorflow_mnist/models/lenet5.uff
    #uff-input-dims=-1,28,28,0 or =3,28,28,0?
    #onnx-file=/usr/src/tensorrt/data/mnist/mnist.onnx
    #int8-calib-file=/usr/src/tensorrt/data/mnist/LegacyCalibrationTable 
    input-dims=1;28;28;0
    uff-input-blob-name=input_1
    output-blob-names=dense_1/Softmax
    #output-blob-names=prob
    enable-dbscan=1
    #bbox-func-name=NvDsInferParseCustomSSD
    #custom-lib-path=/opt/nvidia/deepstream/deepstream-4.0/sources/objectDetector_SSD/nvdsinfer_custom_impl_ssd/libnvdsinfer_custom_impl_ssd.so
	

    [class-attrs-all]
    threshold=0.2
    group-threshold=1
    ## Set eps=0.7 and minBoxes for enable-dbscan=1
    eps=0.2
    #minBoxes=3
    roi-top-offset=0
    roi-bottom-offset=0
    detected-min-w=0
    detected-min-h=0
    detected-max-w=0
    detected-max-h=0

command deepstream-app -c source12_1080p_dec_infer-resnet_tracker_tiled_display_fp16_tx2_teste.txt
with model-color-format = 0 the error is

Creating LL OSD context new
0:00:01.548911553  7467     0x137d5470 INFO                 nvinfer gstnvinfer.cpp:519:gst_nvinfer_logger:<primary_gie_classifier> NvDsInferContext[UID 1]:initialize(): Trying to create engine from model files
0:00:01.788277666  7467     0x137d5470 WARN                 nvinfer gstnvinfer.cpp:515:gst_nvinfer_logger:<primary_gie_classifier> NvDsInferContext[UID 1]:generateTRTModel(): INT8 calibration file not specified. Trying FP16 mode.
0:00:13.845660468  7467     0x137d5470 INFO                 nvinfer gstnvinfer.cpp:519:gst_nvinfer_logger:<primary_gie_classifier> NvDsInferContext[UID 1]:generateTRTModel(): Storing the serialized cuda engine to file at /usr/src/tensorrt/samples/python/end_to_end_tensorflow_mnist/models/lenet5.uff_b8_fp16.engine
0:00:13.893606770  7467     0x137d5470 ERROR                nvinfer gstnvinfer.cpp:511:gst_nvinfer_logger:<primary_gie_classifier> NvDsInferContext[UID 1]:initialize(): RGB/BGR input format specified but network input channels is not 3
0:00:13.899066978  7467     0x137d5470 WARN                 nvinfer gstnvinfer.cpp:692:gst_nvinfer_start:<primary_gie_classifier> error: Failed to create NvDsInferContext instance
0:00:13.899173959  7467     0x137d5470 WARN                 nvinfer gstnvinfer.cpp:692:gst_nvinfer_start:<primary_gie_classifier> error: Config file path: /opt/nvidia/deepstream/deepstream-4.0/samples/configs/deepstream-app/config_infer_primary_nano_teste.txt, NvDsInfer Error: NVDSINFER_CONFIG_FAILED
** ERROR: <main:651>: Failed to set pipeline to PAUSED
Quitting
ERROR from primary_gie_classifier: Failed to create NvDsInferContext instance
Debug info: /dvs/git/dirty/git-master_linux/deepstream/sdk/src/gst-plugins/gst-nvinfer/gstnvinfer.cpp(692): gst_nvinfer_start (): /GstPipeline:pipeline/GstBin:primary_gie_bin/GstNvInfer:primary_gie_classifier:
Config file path: /opt/nvidia/deepstream/deepstream-4.0/samples/configs/deepstream-app/config_infer_primary_nano_teste.txt, NvDsInfer Error: NVDSINFER_CONFIG_FAILED
App run failed

The model is lenet5 in the repository /usr/src/tensorrt/samples/python/end_to_end_tensorflow_mnist

The .pb file was generated from the command:

python3 model.py

2020-02-21 18:38:31.934098: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcudart.so.10.0
WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/tensorflow_core/python/ops/resource_variable_ops.py:1630: calling BaseResourceVariable.__init__ (from tensorflow.python.ops.resource_variable_ops) with constraint is deprecated and will be removed in a future version.
Instructions for updating:
If using Keras pass *_constraint arguments to layers.
Train on 60000 samples
2020-02-21 18:38:42.677882: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcuda.so.1
2020-02-21 18:38:42.692908: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:950] ARM64 does not support NUMA - returning NUMA node zero
2020-02-21 18:38:42.693136: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1618] Found device 0 with properties: 
name: Xavier major: 7 minor: 2 memoryClockRate(GHz): 1.377
pciBusID: 0000:00:00.0
2020-02-21 18:38:42.693240: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcudart.so.10.0
2020-02-21 18:38:42.698725: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcublas.so.10.0
2020-02-21 18:38:42.701888: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcufft.so.10.0
2020-02-21 18:38:42.703251: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcurand.so.10.0
2020-02-21 18:38:42.707923: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcusolver.so.10.0
2020-02-21 18:38:42.711808: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcusparse.so.10.0
2020-02-21 18:38:42.724635: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcudnn.so.7
2020-02-21 18:38:42.725208: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:950] ARM64 does not support NUMA - returning NUMA node zero
2020-02-21 18:38:42.725670: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:950] ARM64 does not support NUMA - returning NUMA node zero
2020-02-21 18:38:42.725856: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1746] Adding visible gpu devices: 0
2020-02-21 18:38:42.776560: W tensorflow/core/platform/profile_utils/cpu_utils.cc:98] Failed to find bogomips in /proc/cpuinfo; cannot determine CPU frequency
2020-02-21 18:38:42.777955: I tensorflow/compiler/xla/service/service.cc:168] XLA service 0x328c91a0 initialized for platform Host (this does not guarantee that XLA will be used). Devices:
2020-02-21 18:38:42.778165: I tensorflow/compiler/xla/service/service.cc:176]   StreamExecutor device (0): Host, Default Version
2020-02-21 18:38:42.906570: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:950] ARM64 does not support NUMA - returning NUMA node zero
2020-02-21 18:38:42.907563: I tensorflow/compiler/xla/service/service.cc:168] XLA service 0x320a6000 initialized for platform CUDA (this does not guarantee that XLA will be used). Devices:
2020-02-21 18:38:42.907717: I tensorflow/compiler/xla/service/service.cc:176]   StreamExecutor device (0): Xavier, Compute Capability 7.2
2020-02-21 18:38:42.908672: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:950] ARM64 does not support NUMA - returning NUMA node zero
2020-02-21 18:38:42.909025: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1618] Found device 0 with properties: 
name: Xavier major: 7 minor: 2 memoryClockRate(GHz): 1.377
pciBusID: 0000:00:00.0
2020-02-21 18:38:42.909142: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcudart.so.10.0
2020-02-21 18:38:42.909221: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcublas.so.10.0
2020-02-21 18:38:42.909279: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcufft.so.10.0
2020-02-21 18:38:42.909332: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcurand.so.10.0
2020-02-21 18:38:42.909387: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcusolver.so.10.0
2020-02-21 18:38:42.909451: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcusparse.so.10.0
2020-02-21 18:38:42.909543: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcudnn.so.7
2020-02-21 18:38:42.909764: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:950] ARM64 does not support NUMA - returning NUMA node zero
2020-02-21 18:38:42.910006: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:950] ARM64 does not support NUMA - returning NUMA node zero
2020-02-21 18:38:42.910129: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1746] Adding visible gpu devices: 0
2020-02-21 18:38:42.910270: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcudart.so.10.0
2020-02-21 18:38:44.373029: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1159] Device interconnect StreamExecutor with strength 1 edge matrix:
2020-02-21 18:38:44.373306: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1165]      0 
2020-02-21 18:38:44.373374: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1178] 0:   N 
2020-02-21 18:38:44.374098: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:950] ARM64 does not support NUMA - returning NUMA node zero
2020-02-21 18:38:44.374509: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:950] ARM64 does not support NUMA - returning NUMA node zero
2020-02-21 18:38:44.374970: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1304] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 5782 MB memory) -> physical GPU (device: 0, name: Xavier, pci bus id: 0000:00:00.0, compute capability: 7.2)
Epoch 1/5
2020-02-21 18:38:48.056166: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcublas.so.10.0
60000/60000 [==============================] - 27s 446us/sample - loss: 0.1995 - acc: 0.9418
Epoch 2/5
60000/60000 [==============================] - 25s 420us/sample - loss: 0.0812 - acc: 0.9749
Epoch 3/5
60000/60000 [==============================] - 25s 420us/sample - loss: 0.0523 - acc: 0.9840
Epoch 4/5
60000/60000 [==============================] - 25s 419us/sample - loss: 0.0361 - acc: 0.9892
Epoch 5/5
60000/60000 [==============================] - 25s 417us/sample - loss: 0.0265 - acc: 0.9920
10000/10000 [==============================] - 3s 270us/sample - loss: 0.0651 - acc: 0.9806
WARNING:tensorflow:From model.py:78: The name tf.keras.backend.get_session is deprecated. Please use tf.compat.v1.keras.backend.get_session instead.

WARNING:tensorflow:From model.py:79: convert_variables_to_constants (from tensorflow.python.framework.graph_util_impl) is deprecated and will be removed in a future version.
Instructions for updating:
Use `tf.compat.v1.graph_util.convert_variables_to_constants`
WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/tensorflow_core/python/framework/graph_util_impl.py:277: extract_sub_graph (from tensorflow.python.framework.graph_util_impl) is deprecated and will be removed in a future version.
Instructions for updating:
Use `tf.compat.v1.graph_util.extract_sub_graph`
WARNING:tensorflow:From model.py:80: remove_training_nodes (from tensorflow.python.framework.graph_util_impl) is deprecated and will be removed in a future version.
Instructions for updating:
Use `tf.compat.v1.graph_util.remove_training_nodes`

then I converted to the uff file using the command in folder /usr/src/tensorrt/samples/python/end_to_end_tensorflow_mnist/models

sudo python3 /usr/lib/python3.6/dist-packages/uff/bin/convert_to_uff.py lenet5.pb lnet5.uff

2020-02-20 14:26:49.647247: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcudart.so.10.0
Loading lenet5.pb
WARNING:tensorflow:From /usr/lib/python3.6/dist-packages/uff/bin/../../uff/converters/tensorflow/conversion_helpers.py:227: The name tf.gfile.GFile is deprecated. Please use tf.io.gfile.GFile instead.

NOTE: UFF has been tested with TensorFlow 1.14.0.
WARNING: The version of TensorFlow installed on this system is not guaranteed to work with UFF.
UFF Version 0.6.5
=== Automatically deduced input nodes ===
[name: "input_1"
op: "Placeholder"
attr {
  key: "dtype"
  value {
    type: DT_FLOAT
  }
}
attr {
  key: "shape"
  value {
    shape {
      dim {
        size: -1
      }
      dim {
        size: 28
      }
      dim {
        size: 28
      }
      dim {
        size: 1
      }
    }
  }
}
]
=========================================

=== Automatically deduced output nodes ===
[name: "dense_1/Softmax"
op: "Softmax"
input: "dense_1/BiasAdd"
attr {
  key: "T"
  value {
    type: DT_FLOAT
  }
}
]
==========================================

Using output node dense_1/Softmax
Converting to UFF graph
DEBUG: convert reshape to flatten node
DEBUG [/usr/lib/python3.6/dist-packages/uff/bin/../../uff/converters/tensorflow/converter.py:96] Marking ['dense_1/Softmax'] as outputs
No. nodes: 13
UFF Output written to lenet5.uff

after then I tested the command in folder /usr/src/tensorrt/bin
./trtexec --uff=/usr/src/tensorrt/samples/python/end_to_end_tensorflow_mnist/models/lenet5.uff --batch=1 --output=dense_1/Softmax --uffInput=input_1,1,28,28 --saveEngine=lenet5.engine --verbose

&&&& RUNNING TensorRT.trtexec # ./trtexec --uff=/usr/src/tensorrt/samples/python/end_to_end_tensorflow_mnist/models/lenet5.uff --batch=1 --output=dense_1/Softmax --uffInput=input_1,1,28,28 --saveEngine=lenet5.engine --verbose
[01/21/2020-18:52:06] [I] === Model Options ===
[01/21/2020-18:52:06] [I] Format: UFF
[01/21/2020-18:52:06] [I] Model: /usr/src/tensorrt/samples/python/end_to_end_tensorflow_mnist/models/lenet5.uff
[01/21/2020-18:52:06] [I] Uff Inputs Layout: NCHW
[01/21/2020-18:52:06] [I] Input: input_1,1,28,28
[01/21/2020-18:52:06] [I] Output: dense_1/Softmax
[01/21/2020-18:52:06] [I] === Build Options ===
[01/21/2020-18:52:06] [I] Max batch: 1
[01/21/2020-18:52:06] [I] Workspace: 16 MB
[01/21/2020-18:52:06] [I] minTiming: 1
[01/21/2020-18:52:06] [I] avgTiming: 8
[01/21/2020-18:52:06] [I] Precision: FP32
[01/21/2020-18:52:06] [I] Calibration: 
[01/21/2020-18:52:06] [I] Safe mode: Disabled
[01/21/2020-18:52:06] [I] Save engine: lenet5.engine
[01/21/2020-18:52:06] [I] Load engine: 
[01/21/2020-18:52:06] [I] Inputs format: fp32:CHW
[01/21/2020-18:52:06] [I] Outputs format: fp32:CHW
[01/21/2020-18:52:06] [I] Input build shapes: model
[01/21/2020-18:52:06] [I] === System Options ===
[01/21/2020-18:52:06] [I] Device: 0
[01/21/2020-18:52:06] [I] DLACore: 
[01/21/2020-18:52:06] [I] Plugins:
[01/21/2020-18:52:06] [I] === Inference Options ===
[01/21/2020-18:52:06] [I] Batch: 1
[01/21/2020-18:52:06] [I] Iterations: 10 (200 ms warm up)
[01/21/2020-18:52:06] [I] Duration: 10s
[01/21/2020-18:52:06] [I] Sleep time: 0ms
[01/21/2020-18:52:06] [I] Streams: 1
[01/21/2020-18:52:06] [I] Spin-wait: Disabled
[01/21/2020-18:52:06] [I] Multithreading: Enabled
[01/21/2020-18:52:06] [I] CUDA Graph: Disabled
[01/21/2020-18:52:06] [I] Skip inference: Disabled
[01/21/2020-18:52:06] [I] Input inference shapes: model
[01/21/2020-18:52:06] [I] === Reporting Options ===
[01/21/2020-18:52:06] [I] Verbose: Enabled
[01/21/2020-18:52:06] [I] Averages: 10 inferences
[01/21/2020-18:52:06] [I] Percentile: 99
[01/21/2020-18:52:06] [I] Dump output: Disabled
[01/21/2020-18:52:06] [I] Profile: Disabled
[01/21/2020-18:52:06] [I] Export timing to JSON file: 
[01/21/2020-18:52:06] [I] Export profile to JSON file: 
[01/21/2020-18:52:06] [I] 
[01/21/2020-18:52:06] [V] [TRT] Plugin Creator registration succeeded - GridAnchor_TRT
[01/21/2020-18:52:06] [V] [TRT] Plugin Creator registration succeeded - GridAnchorRect_TRT
[01/21/2020-18:52:06] [V] [TRT] Plugin Creator registration succeeded - NMS_TRT
[01/21/2020-18:52:06] [V] [TRT] Plugin Creator registration succeeded - Reorg_TRT
[01/21/2020-18:52:06] [V] [TRT] Plugin Creator registration succeeded - Region_TRT
[01/21/2020-18:52:06] [V] [TRT] Plugin Creator registration succeeded - Clip_TRT
[01/21/2020-18:52:06] [V] [TRT] Plugin Creator registration succeeded - LReLU_TRT
[01/21/2020-18:52:06] [V] [TRT] Plugin Creator registration succeeded - PriorBox_TRT
[01/21/2020-18:52:06] [V] [TRT] Plugin Creator registration succeeded - Normalize_TRT
[01/21/2020-18:52:06] [V] [TRT] Plugin Creator registration succeeded - RPROI_TRT
[01/21/2020-18:52:06] [V] [TRT] Plugin Creator registration succeeded - BatchedNMS_TRT
[01/21/2020-18:52:06] [V] [TRT] Plugin Creator registration succeeded - FlattenConcat_TRT
[01/21/2020-18:52:07] [V] [TRT] UFFParser: Parsing input_1[Op: Input].
[01/21/2020-18:52:07] [V] [TRT] UFFParser: input_1 -> [1,28,28]
[01/21/2020-18:52:07] [V] [TRT] UFFParser: Applying order forwarding to: input_1
[01/21/2020-18:52:07] [V] [TRT] UFFParser: Parsing flatten/Reshape[Op: Flatten]. Inputs: input_1
[01/21/2020-18:52:07] [V] [TRT] UFFParser: input_1 -> [1,28,28]
[01/21/2020-18:52:07] [V] [TRT] UFFParser: Applying order forwarding to: flatten/Reshape
[01/21/2020-18:52:07] [V] [TRT] UFFParser: Parsing dense/kernel[Op: Const].
[01/21/2020-18:52:07] [V] [TRT] UFFParser: dense/kernel -> [784,512]
[01/21/2020-18:52:07] [V] [TRT] UFFParser: Applying order forwarding to: dense/kernel
[01/21/2020-18:52:07] [V] [TRT] UFFParser: Parsing dense/MatMul[Op: FullyConnected]. Inputs: flatten/Reshape, dense/kernel
[01/21/2020-18:52:07] [V] [TRT] UFFParser: dense/MatMul -> [512,1,1]
[01/21/2020-18:52:07] [V] [TRT] UFFParser: Applying order forwarding to: dense/MatMul
[01/21/2020-18:52:07] [V] [TRT] UFFParser: Parsing dense/bias[Op: Const].
[01/21/2020-18:52:07] [V] [TRT] UFFParser: dense/bias -> [512]
[01/21/2020-18:52:07] [V] [TRT] UFFParser: Applying order forwarding to: dense/bias
[01/21/2020-18:52:07] [V] [TRT] UFFParser: Parsing dense/BiasAdd[Op: Binary]. Inputs: dense/MatMul, dense/bias
[01/21/2020-18:52:07] [V] [TRT] UFFParser: dense/BiasAdd -> [512,1,1]
[01/21/2020-18:52:07] [V] [TRT] UFFParser: Applying order forwarding to: dense/BiasAdd
[01/21/2020-18:52:07] [V] [TRT] UFFParser: Parsing dense/Relu[Op: Activation]. Inputs: dense/BiasAdd
[01/21/2020-18:52:07] [V] [TRT] UFFParser: dense/Relu -> [512,1,1]
[01/21/2020-18:52:07] [V] [TRT] UFFParser: Applying order forwarding to: dense/Relu
[01/21/2020-18:52:07] [V] [TRT] UFFParser: Parsing dense_1/kernel[Op: Const].
[01/21/2020-18:52:07] [V] [TRT] UFFParser: dense_1/kernel -> [512,10]
[01/21/2020-18:52:07] [V] [TRT] UFFParser: Applying order forwarding to: dense_1/kernel
[01/21/2020-18:52:07] [V] [TRT] UFFParser: Parsing dense_1/MatMul[Op: FullyConnected]. Inputs: dense/Relu, dense_1/kernel
[01/21/2020-18:52:07] [V] [TRT] UFFParser: dense_1/MatMul -> [10,1,1]
[01/21/2020-18:52:07] [V] [TRT] UFFParser: Applying order forwarding to: dense_1/MatMul
[01/21/2020-18:52:07] [V] [TRT] UFFParser: Parsing dense_1/bias[Op: Const].
[01/21/2020-18:52:07] [V] [TRT] UFFParser: dense_1/bias -> [10]
[01/21/2020-18:52:07] [V] [TRT] UFFParser: Applying order forwarding to: dense_1/bias
[01/21/2020-18:52:07] [V] [TRT] UFFParser: Parsing dense_1/BiasAdd[Op: Binary]. Inputs: dense_1/MatMul, dense_1/bias
[01/21/2020-18:52:07] [V] [TRT] UFFParser: dense_1/BiasAdd -> [10,1,1]
[01/21/2020-18:52:07] [V] [TRT] UFFParser: Applying order forwarding to: dense_1/BiasAdd
[01/21/2020-18:52:07] [V] [TRT] UFFParser: Parsing dense_1/Softmax[Op: Softmax]. Inputs: dense_1/BiasAdd
[01/21/2020-18:52:07] [V] [TRT] UFFParser: dense_1/Softmax -> [10,1,1]
[01/21/2020-18:52:07] [V] [TRT] UFFParser: Applying order forwarding to: dense_1/Softmax
[01/21/2020-18:52:07] [V] [TRT] UFFParser: Parsing MarkOutput_0[Op: MarkOutput]. Inputs: dense_1/Softmax
[01/21/2020-18:52:07] [V] [TRT] UFFParser: Applying order forwarding to: MarkOutput_0
[01/21/2020-18:52:07] [V] [TRT] Applying generic optimizations to the graph for inference.
[01/21/2020-18:52:07] [V] [TRT] Original: 11 layers
[01/21/2020-18:52:07] [V] [TRT] After dead-layer removal: 7 layers
[01/21/2020-18:52:07] [V] [TRT] After scale fusion: 7 layers
[01/21/2020-18:52:07] [V] [TRT] Fusing dense/MatMul with dense/BiasAdd
[01/21/2020-18:52:07] [V] [TRT] Fusing dense/MatMul + dense/BiasAdd with dense/Relu
[01/21/2020-18:52:07] [V] [TRT] Fusing dense_1/MatMul with dense_1/BiasAdd
[01/21/2020-18:52:07] [V] [TRT] After vertical fusions: 4 layers
[01/21/2020-18:52:07] [V] [TRT] After final dead-layer removal: 4 layers
[01/21/2020-18:52:07] [V] [TRT] After tensor merging: 4 layers
[01/21/2020-18:52:07] [V] [TRT] After concat removal: 4 layers
[01/21/2020-18:52:07] [V] [TRT] Graph construction and optimization completed in 0.00325652 seconds.
[01/21/2020-18:52:07] [I] [TRT] 
[01/21/2020-18:52:07] [I] [TRT] --------------- Layers running on DLA: 
[01/21/2020-18:52:07] [I] [TRT] 
[01/21/2020-18:52:07] [I] [TRT] --------------- Layers running on GPU: 
[01/21/2020-18:52:07] [I] [TRT] dense/MatMul + dense/BiasAdd + dense/Relu, dense_1/MatMul + dense_1/BiasAdd, dense_1/Softmax_HL_1804289383, dense_1/Softmax, 
[01/21/2020-18:52:12] [V] [TRT] Constructing optimization profile number 0 out of 1
*************** Autotuning format combination: Float(1,28,784,784) -> Float(1,1,1,512) ***************
[01/21/2020-18:52:12] [V] [TRT] --------------- Timing Runner: dense/MatMul + dense/BiasAdd + dense/Relu (CaskFullyConnected)
[01/21/2020-18:52:12] [V] [TRT] dense/MatMul + dense/BiasAdd + dense/Relu (caskFullyConnectedFP32) Set Tactic Name: volta_sgemm_128x32_relu_nn_v1
[01/21/2020-18:52:12] [V] [TRT] Tactic: 8190451076631534509 time 0.278528
[01/21/2020-18:52:12] [V] [TRT] dense/MatMul + dense/BiasAdd + dense/Relu (caskFullyConnectedFP32) Set Tactic Name: volta_sgemm_32x128_relu_nn_v1
[01/21/2020-18:52:12] [V] [TRT] Tactic: 5918215112415401770 time 0.37072
[01/21/2020-18:52:12] [V] [TRT] dense/MatMul + dense/BiasAdd + dense/Relu (caskFullyConnectedFP32) Set Tactic Name: volta_sgemm_64x64_relu_nn_v1
[01/21/2020-18:52:12] [V] [TRT] Tactic: 3514307671120560729 time 0.391168
[01/21/2020-18:52:12] [V] [TRT] dense/MatMul + dense/BiasAdd + dense/Relu (caskFullyConnectedFP32) Set Tactic Name: volta_sgemm_128x128_relu_nn_v1
[01/21/2020-18:52:12] [V] [TRT] Tactic: 3380542644225735979 time 0.682976
[01/21/2020-18:52:12] [V] [TRT] dense/MatMul + dense/BiasAdd + dense/Relu (caskFullyConnectedFP32) Set Tactic Name: volta_sgemm_128x64_relu_nn_v1
[01/21/2020-18:52:12] [V] [TRT] Tactic: 2365066484941477695 time 0.421856
[01/21/2020-18:52:12] [V] [TRT] Fastest Tactic: 8190451076631534509 Time: 0.278528
[01/21/2020-18:52:12] [V] [TRT] --------------- Timing Runner: dense/MatMul + dense/BiasAdd + dense/Relu (CudaFullyConnected)
[01/21/2020-18:52:12] [V] [TRT] Tactic: 0 time 0.125888
[01/21/2020-18:52:12] [V] [TRT] Tactic: 1 time 0.123904
[01/21/2020-18:52:12] [V] [TRT] Fastest Tactic: 1 Time: 0.123904
[01/21/2020-18:52:12] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: CudaFullyConnected Tactic: 1
[01/21/2020-18:52:12] [V] [TRT] 
[01/21/2020-18:52:12] [V] [TRT] *************** Autotuning format combination: Float(1,1,1,512) -> Float(1,1,1,10) ***************
[01/21/2020-18:52:12] [V] [TRT] --------------- Timing Runner: dense_1/MatMul + dense_1/BiasAdd (CaskFullyConnected)
[01/21/2020-18:52:12] [V] [TRT] dense_1/MatMul + dense_1/BiasAdd (caskFullyConnectedFP32) Set Tactic Name: volta_sgemm_128x32_relu_nn_v1
[01/21/2020-18:52:12] [V] [TRT] Tactic: 8190451076631534509 time 0.128992
[01/21/2020-18:52:12] [V] [TRT] dense_1/MatMul + dense_1/BiasAdd (caskFullyConnectedFP32) Set Tactic Name: volta_sgemm_32x128_relu_nn_v1
[01/21/2020-18:52:12] [V] [TRT] Tactic: 5918215112415401770 time 0.133152
[01/21/2020-18:52:12] [V] [TRT] dense_1/MatMul + dense_1/BiasAdd (caskFullyConnectedFP32) Set Tactic Name: volta_sgemm_64x64_relu_nn_v1
[01/21/2020-18:52:12] [V] [TRT] Tactic: 3514307671120560729 time 0.256992
[01/21/2020-18:52:12] [V] [TRT] dense_1/MatMul + dense_1/BiasAdd (caskFullyConnectedFP32) Set Tactic Name: volta_sgemm_128x128_relu_nn_v1
[01/21/2020-18:52:12] [V] [TRT] Tactic: 3380542644225735979 time 0.466976
[01/21/2020-18:52:12] [V] [TRT] dense_1/MatMul + dense_1/BiasAdd (caskFullyConnectedFP32) Set Tactic Name: volta_sgemm_128x64_relu_nn_v1
[01/21/2020-18:52:12] [V] [TRT] Tactic: 2365066484941477695 time 0.249856
[01/21/2020-18:52:12] [V] [TRT] Fastest Tactic: 8190451076631534509 Time: 0.128992
[01/21/2020-18:52:12] [V] [TRT] --------------- Timing Runner: dense_1/MatMul + dense_1/BiasAdd (CudaFullyConnected)
[01/21/2020-18:52:12] [V] [TRT] Tactic: 0 time 0.043008
[01/21/2020-18:52:12] [V] [TRT] Tactic: 1 time 0.033728
[01/21/2020-18:52:12] [V] [TRT] Fastest Tactic: 1 Time: 0.033728
[01/21/2020-18:52:12] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: CudaFullyConnected Tactic: 1
[01/21/2020-18:52:12] [V] [TRT] 
[01/21/2020-18:52:12] [V] [TRT] *************** Autotuning format combination: Float(1,1,1,10) -> Float(1,1,1,10) ***************
[01/21/2020-18:52:12] [V] [TRT] --------------- Timing Runner: dense_1/Softmax_HL_1804289383 (SoftMax)
[01/21/2020-18:52:12] [V] [TRT] Tactic: 1001 time 0.020448
[01/21/2020-18:52:12] [V] [TRT] Fastest Tactic: 1001 Time: 0.020448
[01/21/2020-18:52:12] [V] [TRT] >>>>>>>>>>>>>>> Chose Runner Type: SoftMax Tactic: 1001
[01/21/2020-18:52:12] [V] [TRT] 
[01/21/2020-18:52:12] [V] [TRT] *************** Autotuning format combination: Float(1,1,1,10) -> Float(1,10,10,10) ***************
[01/21/2020-18:52:12] [V] [TRT] --------------- Timing Runner: dense_1/Softmax (Shuffle)
[01/21/2020-18:52:12] [V] [TRT] Tactic: 0 is the only option, timing skipped
[01/21/2020-18:52:12] [V] [TRT] Fastest Tactic: 0 Time: 0
[01/21/2020-18:52:12] [V] [TRT] Formats and tactics selection completed in 0.255097 seconds.
[01/21/2020-18:52:12] [V] [TRT] After reformat layers: 4 layers
[01/21/2020-18:52:12] [V] [TRT] Block size 16777216
[01/21/2020-18:52:12] [V] [TRT] Block size 2048
[01/21/2020-18:52:12] [V] [TRT] Block size 512
[01/21/2020-18:52:12] [V] [TRT] Total Activation Memory: 16779776
[01/21/2020-18:52:12] [I] [TRT] Detected 1 inputs and 1 output network tensors.
[01/21/2020-18:52:12] [V] [TRT] Engine generation completed in 5.13601 seconds.
[01/21/2020-18:52:12] [V] [TRT] Engine Layer Information:
[01/21/2020-18:52:12] [V] [TRT] Layer: dense/MatMul + dense/BiasAdd + dense/Relu (FullyConnected), Tactic: 1, input_1[Float(1,28,28)] -> dense/Relu[Float(512,1,1)]
[01/21/2020-18:52:12] [V] [TRT] Layer: dense_1/MatMul + dense_1/BiasAdd (FullyConnected), Tactic: 1, dense/Relu[Float(512,1,1)] -> dense_1/BiasAdd[Float(10,1,1)]
[01/21/2020-18:52:12] [V] [TRT] Layer: dense_1/Softmax_HL_1804289383 (SoftMax), Tactic: 1001, dense_1/BiasAdd[Float(10,1,1)] -> dense_1/Softmax_HL_1804289383[Float(10,1,1)]
[01/21/2020-18:52:12] [V] [TRT] Layer: dense_1/Softmax (Shuffle), Tactic: 0, dense_1/Softmax_HL_1804289383[Float(10,1,1)] -> dense_1/Softmax[Float(1,1,10)]
[01/21/2020-18:52:13] [I] Average over 10 runs is 0.468269 ms (host walltime is 0.547745 ms, 99% percentile time is 0.795712).
[01/21/2020-18:52:13] [I] Average over 10 runs is 0.320851 ms (host walltime is 0.362989 ms, 99% percentile time is 0.567168).
[01/21/2020-18:52:13] [I] Average over 10 runs is 0.296128 ms (host walltime is 0.334411 ms, 99% percentile time is 0.365568).
[01/21/2020-18:52:13] [I] Average over 10 runs is 0.295936 ms (host walltime is 0.336683 ms, 99% percentile time is 0.350208).
[01/21/2020-18:52:13] [I] Average over 10 runs is 0.29145 ms (host walltime is 0.327396 ms, 99% percentile time is 0.33888).
[01/21/2020-18:52:13] [I] Average over 10 runs is 0.456618 ms (host walltime is 0.506729 ms, 99% percentile time is 1.90154).
[01/21/2020-18:52:13] [I] Average over 10 runs is 0.158112 ms (host walltime is 1.03403 ms, 99% percentile time is 0.164896).
[01/21/2020-18:52:13] [I] Average over 10 runs is 0.281194 ms (host walltime is 1.05704 ms, 99% percentile time is 1.16944).
[01/21/2020-18:52:13] [I] Average over 10 runs is 0.395085 ms (host walltime is 0.901245 ms, 99% percentile time is 1.65485).
[01/21/2020-18:52:13] [I] Average over 10 runs is 0.261875 ms (host walltime is 0.294909 ms, 99% percentile time is 0.31872).
&&&& PASSED TensorRT.trtexec # ./trtexec --uff=/usr/src/tensorrt/samples/python/end_to_end_tensorflow_mnist/models/lenet5.uff --batch=1 --output=dense_1/Softmax --uffInput=input_1,1,28,28 --saveEngine=lenet5.engine --verbose

The lenet5.engine was generate successfully

deepstream-app --version-all

deepstream-app version 4.0.2
DeepStreamSDK 4.0.2
CUDA Driver Version: 10.0
CUDA Runtime Version: 10.0
TensorRT Version: 6.0
cuDNN Version: 7.6
libNVWarp360 Version: 2.0.0d5

how can i resolve this error? Do I need an extra code to be able to run lenet5 deepstream?

+1

I’m getting the same error.

+1

I am having the same problem.

Hi lucasssss,
for lenet5, since it’s single channel/gray input, so you should set model-color-format = 2 .

And, with “model-color-format = 2”, what failure log did you see?

File: source12_1080p_dec_infer-resnet_tracker_tiled_display_fp16_tx2_teste.txt

# Copyright (c) 2019 NVIDIA Corporation.  All rights reserved.
#
# NVIDIA Corporation and its licensors retain all intellectual property
# and proprietary rights in and to this software, related documentation
# and any modifications thereto.  Any use, reproduction, disclosure or
# distribution of this software and related documentation without an express
# license agreement from NVIDIA Corporation is strictly prohibited.

[application]
enable-perf-measurement=1
perf-measurement-interval-sec=5
#gie-kitti-output-dir=streamscl

[tiled-display]
enable=0
rows=4
columns=3
width=1280
height=720
gpu-id=0
#(0): nvbuf-mem-default - Default memory allocated, specific to particular platform
#(1): nvbuf-mem-cuda-pinned - Allocate Pinned/Host cuda memory, applicable for Tesla
#(2): nvbuf-mem-cuda-device - Allocate Device cuda memory, applicable for Tesla
#(3): nvbuf-mem-cuda-unified - Allocate Unified cuda memory, applicable for Tesla
#(4): nvbuf-mem-surface-array - Allocate Surface Array memory, applicable for Jetson
nvbuf-memory-type=0

[source0]
enable=1
#Type - 1=CameraV4L2 2=URI 3=MultiURI 4=RTSP
type=3
uri=file://../../streams/sample_1080p_h264.mp4
num-sources=12
#drop-frame-interval=2
gpu-id=0
# (0): memtype_device   - Memory type Device
# (1): memtype_pinned   - Memory type Host Pinned
# (2): memtype_unified  - Memory type Unified
cudadec-memtype=0

[sink0]
enable=1
#Type - 1=FakeSink 2=EglSink 3=File
type=5
sync=1
source-id=0
gpu-id=0
qos=0
nvbuf-memory-type=0
overlay-id=1

[sink1]
enable=0
type=3
#1=mp4 2=mkv
container=1
#1=h264 2=h265
codec=1
sync=0
#iframeinterval=10
bitrate=2000000
output-file=out.mp4
source-id=0

[sink2]
enable=0
#Type - 1=FakeSink 2=EglSink 3=File 4=RTSPStreaming
type=4
#1=h264 2=h265
codec=1
sync=0
bitrate=4000000
# set below properties in case of RTSPStreaming
rtsp-port=8554
udp-port=5400

[osd]
enable=1
gpu-id=0
border-width=1
text-size=15
text-color=1;1;1;1;
text-bg-color=0.3;0.3;0.3;1
font=Serif
show-clock=0
clock-x-offset=800
clock-y-offset=820
clock-text-size=12
clock-color=1;0;0;0
nvbuf-memory-type=0

[streammux]
gpu-id=0
##Boolean property to inform muxer that sources are live
live-source=0
batch-size=12
##time out in usec, to wait after the first buffer is available
##to push the batch even if the complete batch is not formed
batched-push-timeout=40000
## Set muxer output width and height
width=1200
height=1200
##Enable to maintain aspect ratio wrt source, and allow black borders, works
##along with width, height properties
enable-padding=0
nvbuf-memory-type=0


# config-file property is mandatory for any gie section.
# Other properties are optional and if set will override the properties set in
# the infer config file.
[primary-gie]
##enable=1
#gpu-id=0
#model-engine-file=/usr/src/tensorrt/bin/lenet5.engine
#batch-size=12
#Required by the app for OSD, not a plugin property
#bbox-border-color0=1;0;0;1
#bbox-border-color1=0;1;1;1
#bbox-border-color2=0;0;1;1
#bbox-border-color3=0;1;0;1
#interval=4
#gie-unique-id=1
#nvbuf-memory-type=0
enable=1
gpu-id=0
bbox-border-color0=1;0;0;1
bbox-border-color1=0;1;1;1
bbox-border-color2=0;0;1;1
bbox-border-color3=0;1;0;1
interval=4
gie-unique-id=1
nvbuf-memory-type=0
config-file=config_infer_primary_nano_teste.txt

[tracker]
enable=0
tracker-width=480
tracker-height=272
#ll-lib-file=/opt/nvidia/deepstream/deepstream-4.0/lib/libnvds_mot_iou.so
ll-lib-file=/opt/nvidia/deepstream/deepstream-4.0/lib/libnvds_mot_klt.so
#ll-config-file required for IOU only
#ll-config-file=iou_config.txt
gpu-id=0

[tests]
file-loop=0

File: config_infer_primary_nano_teste.txt

# Copyright (c) 2018 NVIDIA Corporation.  All rights reserved.
#
# NVIDIA Corporation and its licensors retain all intellectual property
# and proprietary rights in and to this software, related documentation
# and any modifications thereto.  Any use, reproduction, disclosure or
# distribution of this software and related documentation without an express
# license agreement from NVIDIA Corporation is strictly prohibited.

# Following properties are mandatory when engine files are not specified:
#   int8-calib-file(Only in INT8)
#   Caffemodel mandatory properties: model-file, proto-file, output-blob-names
#   UFF: uff-file, input-dims, uff-input-blob-name, output-blob-names
#   ONNX: onnx-file
#
# Mandatory properties for detectors:
#   num-detected-classes
#
# Optional properties for detectors:
#   enable-dbscan(Default=false), interval(Primary mode only, Default=0)
#   custom-lib-path,
#   parse-bbox-func-name
#
# Mandatory properties for classifiers:
#   classifier-threshold, is-classifier
#
# Optional properties for classifiers:
#   classifier-async-mode(Secondary mode only, Default=false)
#
# Optional properties in secondary mode:
#   operate-on-gie-id(Default=0), operate-on-class-ids(Defaults to all classes),
#   input-object-min-width, input-object-min-height, input-object-max-width,
#   input-object-max-height
#
# Following properties are always recommended:
#   batch-size(Default=1)
#
# Other optional properties:
#   net-scale-factor(Default=1), network-mode(Default=0 i.e FP32),
#   model-color-format(Default=0 i.e. RGB) model-engine-file, labelfile-path,
#   mean-file, gie-unique-id(Default=0), offsets, gie-mode (Default=1 i.e. primary),
#   custom-lib-path, network-mode(Default=0 i.e FP32)
#
# The values in the config file are overridden by values set through GObject
# properties.

    [property]
    gpu-id=0
    net-scale-factor=0.0039215697906911373
    batch-size=8
    process-mode=1
    model-color-format=2
    ## 0=FP32, 1=INT8, 2=FP16 mode
    network-mode=1
    num-detected-classes=10
    interval=0
    gie-unique-id=1
    #uff-file=/usr/src/tensorrt/samples/python/uff_custom_plugin/models/trained_lenet5.uff
    #uff-file=../../models/lenet5.uff
    uff-file=/usr/src/tensorrt/samples/python/end_to_end_tensorflow_mnist/models/lenet5.uff
    #uff-input-dims=-1,28,28,0 or =3,28,28,0?
    #onnx-file=/usr/src/tensorrt/data/mnist/mnist.onnx
    #int8-calib-file=/usr/src/tensorrt/data/mnist/LegacyCalibrationTable 
    input-dims=1;28;28;0
    uff-input-blob-name=input_1
    output-blob-names=dense_1/Softmax
    #output-blob-names=prob
    enable-dbscan=0
    #bbox-func-name=NvDsInferParseCustomSSD
    #custom-lib-path=/opt/nvidia/deepstream/deepstream-4.0/sources/objectDetector_SSD/nvdsinfer_custom_impl_ssd/libnvdsinfer_custom_impl_ssd.so
	

    [class-attrs-all]
    threshold=0.2
    group-threshold=1
    ## Set eps=0.7 and minBoxes for enable-dbscan=1
    eps=0.2
    #minBoxes=3
    roi-top-offset=0
    roi-bottom-offset=0
    detected-min-w=0
    detected-min-h=0
    detected-max-w=0
    detected-max-h=0

Command: deepstream-app -c source12_1080p_dec_infer-resnet_tracker_tiled_display_fp16_tx2_teste.txt with model-color-format = 2
console log:

Creating LL OSD context new
0:00:01.612466173 10595     0x13f8a670 INFO                 nvinfer gstnvinfer.cpp:519:gst_nvinfer_logger:<primary_gie_classifier> NvDsInferContext[UID 1]:initialize(): Trying to create engine from model files
0:00:01.848463807 10595     0x13f8a670 WARN                 nvinfer gstnvinfer.cpp:515:gst_nvinfer_logger:<primary_gie_classifier> NvDsInferContext[UID 1]:generateTRTModel(): INT8 calibration file not specified. Trying FP16 mode.
0:00:12.192338983 10595     0x13f8a670 INFO                 nvinfer gstnvinfer.cpp:519:gst_nvinfer_logger:<primary_gie_classifier> NvDsInferContext[UID 1]:generateTRTModel(): Storing the serialized cuda engine to file at /usr/src/tensorrt/samples/python/end_to_end_tensorflow_mnist/models/lenet5.uff_b8_fp16.engine

Runtime commands:
	h: Print this help
	q: Quit

	p: Pause
	r: Resume


**PERF: FPS 0 (Avg)	FPS 1 (Avg)	FPS 2 (Avg)	FPS 3 (Avg)	FPS 4 (Avg)	FPS 5 (Avg)	FPS 6 (Avg)	FPS 7 (Avg)	FPS 8 (Avg)	FPS 9 (Avg)	FPS 10 (Avg)	FPS 11 (Avg)	
**PERF: 0.00 (0.00)	0.00 (0.00)	0.00 (0.00)	0.00 (0.00)	0.00 (0.00)	0.00 (0.00)	0.00 (0.00)	0.00 (0.00)	0.00 (0.00)	0.00 (0.00)	0.00 (0.00)	0.00 (0.00)	
** INFO: <bus_callback:189>: Pipeline ready

Opening in BLOCKING MODE 
Opening in BLOCKING MODE 
Opening in BLOCKING MODE 
Opening in BLOCKING MODE 
NvMMLiteOpen : Block : BlockType = 261 
Opening in BLOCKING MODE 
NvMMLiteOpen : Block : BlockType = 261 
NvMMLiteOpen : Block : BlockType = 261 
Opening in BLOCKING MODE 
NvMMLiteOpen : Block : BlockType = 261 
NVMEDIA: Reading vendor.tegra.display-size : status: 6 
NVMEDIA: Reading vendor.tegra.display-size : status: 6 
NvMMLiteBlockCreate : Block : BlockType = 261 
NVMEDIA: Reading vendor.tegra.display-size : status: 6 
NVMEDIA: Reading vendor.tegra.display-size : status: 6 
NvMMLiteOpen : Block : BlockType = 261 
Opening in BLOCKING MODE 
NVMEDIA: Reading vendor.tegra.display-size : status: 6 
NvMMLiteOpen : Block : BlockType = 261 
NVMEDIA: Reading vendor.tegra.display-size : status: 6 
Opening in BLOCKING MODE 
NvMMLiteOpen : Block : BlockType = 261 
NvMMLiteOpen : Block : BlockType = 261 
NVMEDIA: Reading vendor.tegra.display-size : status: 6 
NvMMLiteBlockCreate : Block : BlockType = 261 
NvMMLiteBlockCreate : Block : BlockType = 261 
NvMMLiteBlockCreate : Block : BlockType = 261 
Opening in BLOCKING MODE 
Opening in BLOCKING MODE 
Opening in BLOCKING MODE 
Opening in BLOCKING MODE 
NvMMLiteOpen : Block : BlockType = 261 
NvMMLiteOpen : Block : BlockType = 261 
NvMMLiteBlockCreate : Block : BlockType = 261 
NvMMLiteOpen : Block : BlockType = 261 
NvMMLiteBlockCreate : Block : BlockType = 261 
NvMMLiteBlockCreate : Block : BlockType = 261 
NvMMLiteOpen : Block : BlockType = 261 
NVMEDIA: Reading vendor.tegra.display-size : status: 6 
NvMMLiteBlockCreate : Block : BlockType = 261 
NVMEDIA: Reading vendor.tegra.display-size : status: 6 
NvMMLiteBlockCreate : Block : BlockType = 261 
NVMEDIA: Reading vendor.tegra.display-size : status: 6 
NVMEDIA: Reading vendor.tegra.display-size : status: 6 
NvMMLiteBlockCreate : Block : BlockType = 261 
NvMMLiteBlockCreate : Block : BlockType = 261 
NVMEDIA: Reading vendor.tegra.display-size : status: 6 
NvMMLiteBlockCreate : Block : BlockType = 261 
Creating LL OSD context new
** INFO: <bus_callback:175>: Pipeline running

0:00:13.406614393 10595     0x13968320 WARN                 nvinfer gstnvinfer.cpp:1149:convert_batch_and_push_to_input_thread:<primary_gie_classifier> error: NvBufSurfTransform failed with error -2 while converting buffer
ERROR from primary_gie_classifier: NvBufSurfTransform failed with error -2 while converting buffer
Debug info: /dvs/git/dirty/git-master_linux/deepstream/sdk/src/gst-plugins/gst-nvinfer/gstnvinfer.cpp(1149): convert_batch_and_push_to_input_thread (): /GstPipeline:pipeline/GstBin:primary_gie_bin/GstNvInfer:primary_gie_classifier
ERROR from qtdemux2: Internal data stream error.
Debug info: qtdemux.c(6073): gst_qtdemux_loop (): /GstPipeline:pipeline/GstBin:multi_src_bin/GstBin:src_sub_bin1/GstURIDecodeBin:src_elem/GstDecodeBin:decodebin1/GstQTDemux:qtdemux2:
streaming stopped, reason error (-5)
ERROR from qtdemux4: Internal data stream error.
Debug info: qtdemux.c(6073): gst_qtdemux_loop (): /GstPipeline:pipeline/GstBin:multi_src_bin/GstBin:src_sub_bin2/GstURIDecodeBin:src_elem/GstDecodeBin:decodebin2/GstQTDemux:qtdemux4:
streaming stopped, reason error (-5)
ERROR from qtdemux9: Internal data stream error.
Debug info: qtdemux.c(6073): gst_qtdemux_loop (): /GstPipeline:pipeline/GstBin:multi_src_bin/GstBin:src_sub_bin8/GstURIDecodeBin:src_elem/GstDecodeBin:decodebin8/GstQTDemux:qtdemux9:
streaming stopped, reason error (-5)
Quitting
ERROR from qtdemux6: Internal data stream error.
Debug info: qtdemux.c(6073): gst_qtdemux_loop (): /GstPipeline:pipeline/GstBin:multi_src_bin/GstBin:src_sub_bin5/GstURIDecodeBin:src_elem/GstDecodeBin:decodebin5/GstQTDemux:qtdemux6:
streaming stopped, reason error (-5)
ERROR from qtdemux5: Internal data stream error.
Debug info: qtdemux.c(6073): gst_qtdemux_loop (): /GstPipeline:pipeline/GstBin:multi_src_bin/GstBin:src_sub_bin6/GstURIDecodeBin:src_elem/GstDecodeBin:decodebin6/GstQTDemux:qtdemux5:
streaming stopped, reason error (-5)
ERROR from qtdemux1: Internal data stream error.
Debug info: qtdemux.c(6073): gst_qtdemux_loop (): /GstPipeline:pipeline/GstBin:multi_src_bin/GstBin:src_sub_bin4/GstURIDecodeBin:src_elem/GstDecodeBin:decodebin4/GstQTDemux:qtdemux1:
streaming stopped, reason error (-5)
ERROR from qtdemux0: Internal data stream error.
Debug info: qtdemux.c(6073): gst_qtdemux_loop (): /GstPipeline:pipeline/GstBin:multi_src_bin/GstBin:src_sub_bin3/GstURIDecodeBin:src_elem/GstDecodeBin:decodebin3/GstQTDemux:qtdemux0:
streaming stopped, reason error (-5)
ERROR from qtdemux7: Internal data stream error.
Debug info: qtdemux.c(6073): gst_qtdemux_loop (): /GstPipeline:pipeline/GstBin:multi_src_bin/GstBin:src_sub_bin7/GstURIDecodeBin:src_elem/GstDecodeBin:decodebin7/GstQTDemux:qtdemux7:
streaming stopped, reason error (-5)
ERROR from qtdemux8: Internal data stream error.
Debug info: qtdemux.c(6073): gst_qtdemux_loop (): /GstPipeline:pipeline/GstBin:multi_src_bin/GstBin:src_sub_bin10/GstURIDecodeBin:src_elem/GstDecodeBin:decodebin10/GstQTDemux:qtdemux8:
streaming stopped, reason error (-5)
ERROR from multiqueue9: Internal data stream error.
Debug info: gstmultiqueue.c(2357): gst_multi_queue_sink_event (): /GstPipeline:pipeline/GstBin:multi_src_bin/GstBin:src_sub_bin10/GstURIDecodeBin:src_elem/GstDecodeBin:decodebin10/GstMultiQueue:multiqueue9:
streaming stopped, reason error (-5)
ERROR from qtdemux3: Internal data stream error.
Debug info: qtdemux.c(6073): gst_qtdemux_loop (): /GstPipeline:pipeline/GstBin:multi_src_bin/GstBin:src_sub_bin0/GstURIDecodeBin:src_elem/GstDecodeBin:decodebin0/GstQTDemux:qtdemux3:
streaming stopped, reason error (-5)
ERROR from qtdemux10: Internal data stream error.
Debug info: qtdemux.c(6073): gst_qtdemux_loop (): /GstPipeline:pipeline/GstBin:multi_src_bin/GstBin:src_sub_bin9/GstURIDecodeBin:src_elem/GstDecodeBin:decodebin9/GstQTDemux:qtdemux10:
streaming stopped, reason error (-5)
App run failed

My error: ERROR from primary_gie_classifier: NvBufSurfTransform failed with error -2 while converting buffer

Hi all,
Thanks for reporting this issue!
It’s a bug of DeepStream, currently, streammux output width and height can’t be 16 times of larger than the width & height of network resolution . We will fix it ASAP.
So, as a workaround for now, you could change the width and height from 1200 to 448 (=28 x 16) in the streammux node.

[streammux]
gpu-id=0
##Boolean property to inform muxer that sources are live
live-source=0
batch-size=12
##time out in usec, to wait after the first buffer is available
##to push the batch even if the complete batch is not formed
batched-push-timeout=40000

Set muxer output width and height

width=1200
height=1200

Hi, mchi.

The error has changed.

My file: source12_1080p_dec_infer-resnet_tracker_tiled_display_fp16_tx2_teste.txt

# Copyright (c) 2019 NVIDIA Corporation.  All rights reserved.
#
# NVIDIA Corporation and its licensors retain all intellectual property
# and proprietary rights in and to this software, related documentation
# and any modifications thereto.  Any use, reproduction, disclosure or
# distribution of this software and related documentation without an express
# license agreement from NVIDIA Corporation is strictly prohibited.

[application]
enable-perf-measurement=1
perf-measurement-interval-sec=5
#gie-kitti-output-dir=streamscl

[tiled-display]
enable=0
rows=4
columns=3
width=1280
height=720
gpu-id=0
#(0): nvbuf-mem-default - Default memory allocated, specific to particular platform
#(1): nvbuf-mem-cuda-pinned - Allocate Pinned/Host cuda memory, applicable for Tesla
#(2): nvbuf-mem-cuda-device - Allocate Device cuda memory, applicable for Tesla
#(3): nvbuf-mem-cuda-unified - Allocate Unified cuda memory, applicable for Tesla
#(4): nvbuf-mem-surface-array - Allocate Surface Array memory, applicable for Jetson
nvbuf-memory-type=0

[source0]
enable=1
#Type - 1=CameraV4L2 2=URI 3=MultiURI 4=RTSP
type=3
uri=file://../../streams/sample_1080p_h264.mp4
num-sources=12
#drop-frame-interval=2
gpu-id=0
# (0): memtype_device   - Memory type Device
# (1): memtype_pinned   - Memory type Host Pinned
# (2): memtype_unified  - Memory type Unified
cudadec-memtype=0

[sink0]
enable=1
#Type - 1=FakeSink 2=EglSink 3=File
type=5
sync=1
source-id=0
gpu-id=0
qos=0
nvbuf-memory-type=0
overlay-id=1

[sink1]
enable=0
type=3
#1=mp4 2=mkv
container=1
#1=h264 2=h265
codec=1
sync=0
#iframeinterval=10
bitrate=2000000
output-file=out.mp4
source-id=0

[sink2]
enable=0
#Type - 1=FakeSink 2=EglSink 3=File 4=RTSPStreaming
type=4
#1=h264 2=h265
codec=1
sync=0
bitrate=4000000
# set below properties in case of RTSPStreaming
rtsp-port=8554
udp-port=5400

[osd]
enable=1
gpu-id=0
border-width=1
text-size=15
text-color=1;1;1;1;
text-bg-color=0.3;0.3;0.3;1
font=Serif
show-clock=0
clock-x-offset=800
clock-y-offset=820
clock-text-size=12
clock-color=1;0;0;0
nvbuf-memory-type=0

[streammux]
gpu-id=0
##Boolean property to inform muxer that sources are live
live-source=0
batch-size=12
##time out in usec, to wait after the first buffer is available
##to push the batch even if the complete batch is not formed
batched-push-timeout=40000
## Set muxer output width and height
#width=1200
#height=1200
width=448
height=448
##Enable to maintain aspect ratio wrt source, and allow black borders, works
##along with width, height properties
enable-padding=0
nvbuf-memory-type=0


# config-file property is mandatory for any gie section.
# Other properties are optional and if set will override the properties set in
# the infer config file.
[primary-gie]
##enable=1
#gpu-id=0
#model-engine-file=/usr/src/tensorrt/bin/lenet5.engine
#batch-size=12
#Required by the app for OSD, not a plugin property
#bbox-border-color0=1;0;0;1
#bbox-border-color1=0;1;1;1
#bbox-border-color2=0;0;1;1
#bbox-border-color3=0;1;0;1
#interval=4
#gie-unique-id=1
#nvbuf-memory-type=0
enable=1
gpu-id=0
bbox-border-color0=1;0;0;1
bbox-border-color1=0;1;1;1
bbox-border-color2=0;0;1;1
bbox-border-color3=0;1;0;1
interval=4
gie-unique-id=1
nvbuf-memory-type=0
config-file=config_infer_primary_nano_teste.txt

[tracker]
enable=0
tracker-width=480
tracker-height=272
#ll-lib-file=/opt/nvidia/deepstream/deepstream-4.0/lib/libnvds_mot_iou.so
ll-lib-file=/opt/nvidia/deepstream/deepstream-4.0/lib/libnvds_mot_klt.so
#ll-config-file required for IOU only
#ll-config-file=iou_config.txt
gpu-id=0

[tests]
file-loop=0

My file: config_infer_primary_nano_teste.txt

# Copyright (c) 2018 NVIDIA Corporation.  All rights reserved.
#
# NVIDIA Corporation and its licensors retain all intellectual property
# and proprietary rights in and to this software, related documentation
# and any modifications thereto.  Any use, reproduction, disclosure or
# distribution of this software and related documentation without an express
# license agreement from NVIDIA Corporation is strictly prohibited.

# Following properties are mandatory when engine files are not specified:
#   int8-calib-file(Only in INT8)
#   Caffemodel mandatory properties: model-file, proto-file, output-blob-names
#   UFF: uff-file, input-dims, uff-input-blob-name, output-blob-names
#   ONNX: onnx-file
#
# Mandatory properties for detectors:
#   num-detected-classes
#
# Optional properties for detectors:
#   enable-dbscan(Default=false), interval(Primary mode only, Default=0)
#   custom-lib-path,
#   parse-bbox-func-name
#
# Mandatory properties for classifiers:
#   classifier-threshold, is-classifier
#
# Optional properties for classifiers:
#   classifier-async-mode(Secondary mode only, Default=false)
#
# Optional properties in secondary mode:
#   operate-on-gie-id(Default=0), operate-on-class-ids(Defaults to all classes),
#   input-object-min-width, input-object-min-height, input-object-max-width,
#   input-object-max-height
#
# Following properties are always recommended:
#   batch-size(Default=1)
#
# Other optional properties:
#   net-scale-factor(Default=1), network-mode(Default=0 i.e FP32),
#   model-color-format(Default=0 i.e. RGB) model-engine-file, labelfile-path,
#   mean-file, gie-unique-id(Default=0), offsets, gie-mode (Default=1 i.e. primary),
#   custom-lib-path, network-mode(Default=0 i.e FP32)
#
# The values in the config file are overridden by values set through GObject
# properties.

    [property]
    gpu-id=0
    net-scale-factor=0.0039215697906911373
    batch-size=8
    process-mode=1
    model-color-format=2
    ## 0=FP32, 1=INT8, 2=FP16 mode
    network-mode=0
    num-detected-classes=10
    interval=0
    gie-unique-id=1
    #uff-file=/usr/src/tensorrt/samples/python/uff_custom_plugin/models/trained_lenet5.uff
    #uff-file=../../models/lenet5.uff
    uff-file=/usr/src/tensorrt/samples/python/end_to_end_tensorflow_mnist/models/lenet5.uff
    #uff-input-dims=-1,28,28,0 or =3,28,28,0?
    #onnx-file=/usr/src/tensorrt/data/mnist/mnist.onnx
    #int8-calib-file=/usr/src/tensorrt/data/mnist/LegacyCalibrationTable 
    input-dims=1;28;28;0
    uff-input-blob-name=input_1
    #output-blob-names=dense_1/Softmax
    output-blob-names=dense_1/Softmax
    #output-blob-names=prob
    #enable-dbscan=0
    #bbox-func-name=NvDsInferParseCustomSSD
    #custom-lib-path=/opt/nvidia/deepstream/deepstream-4.0/sources/objectDetector_SSD/nvdsinfer_custom_impl_ssd/libnvdsinfer_custom_impl_ssd.so
	

    [class-attrs-all]
    threshold=0.2
    group-threshold=1
    ## Set eps=0.7 and minBoxes for enable-dbscan=1
    eps=0.2
    #minBoxes=3
    roi-top-offset=0
    roi-bottom-offset=0
    detected-min-w=0
    detected-min-h=0
    detected-max-w=0
    detected-max-h=0

My command: deepstream-app -c source12_1080p_dec_infer-resnet_tracker_tiled_display_fp16_tx2_teste.txt
terminal log:

Creating LL OSD context new
0:00:01.524100440 19755      0xdeb3a70 INFO                 nvinfer gstnvinfer.cpp:519:gst_nvinfer_logger:<primary_gie_classifier> NvDsInferContext[UID 1]:initialize(): Trying to create engine from model files
0:00:07.356594947 19755      0xdeb3a70 INFO                 nvinfer gstnvinfer.cpp:519:gst_nvinfer_logger:<primary_gie_classifier> NvDsInferContext[UID 1]:generateTRTModel(): Storing the serialized cuda engine to file at /usr/src/tensorrt/samples/python/end_to_end_tensorflow_mnist/models/lenet5.uff_b8_fp32.engine

Runtime commands:
	h: Print this help
	q: Quit

	p: Pause
	r: Resume


**PERF: FPS 0 (Avg)	FPS 1 (Avg)	FPS 2 (Avg)	FPS 3 (Avg)	FPS 4 (Avg)	FPS 5 (Avg)	FPS 6 (Avg)	FPS 7 (Avg)	FPS 8 (Avg)	FPS 9 (Avg)	FPS 10 (Avg)	FPS 11 (Avg)	
**PERF: 0.00 (0.00)	0.00 (0.00)	0.00 (0.00)	0.00 (0.00)	0.00 (0.00)	0.00 (0.00)	0.00 (0.00)	0.00 (0.00)	0.00 (0.00)	0.00 (0.00)	0.00 (0.00)	0.00 (0.00)	
** INFO: <bus_callback:189>: Pipeline ready

Opening in BLOCKING MODE 
Opening in BLOCKING MODE 
Opening in BLOCKING MODE 
Opening in BLOCKING MODE 
Opening in BLOCKING MODE 
Opening in BLOCKING MODE 
Opening in BLOCKING MODE 
NvMMLiteOpen : Block : BlockType = 261 
Opening in BLOCKING MODE 
NvMMLiteOpen : Block : BlockType = 261 
Opening in BLOCKING MODE 
Opening in BLOCKING MODE 
NvMMLiteOpen : Block : BlockType = 261 
NvMMLiteOpen : Block : BlockType = 261 
NvMMLiteOpen : Block : BlockType = 261 
Opening in BLOCKING MODE 
NvMMLiteOpen : Block : BlockType = 261 
NVMEDIA: Reading vendor.tegra.display-size : status: 6 
NVMEDIA: Reading vendor.tegra.display-size : status: 6 
NVMEDIA: Reading vendor.tegra.display-size : status: 6 
NVMEDIA: Reading vendor.tegra.display-size : status: 6 
NVMEDIA: Reading vendor.tegra.display-size : status: 6 
NvMMLiteBlockCreate : Block : BlockType = 261 
NVMEDIA: Reading vendor.tegra.display-size : status: 6 
NvMMLiteOpen : Block : BlockType = 261 
NvMMLiteBlockCreate : Block : BlockType = 261 
NvMMLiteOpen : Block : BlockType = 261 
NvMMLiteBlockCreate : Block : BlockType = 261 
NvMMLiteOpen : Block : BlockType = 261 
NvMMLiteOpen : Block : BlockType = 261 
NvMMLiteOpen : Block : BlockType = 261 
NvMMLiteBlockCreate : Block : BlockType = 261 
Opening in BLOCKING MODE 
NVMEDIA: Reading vendor.tegra.display-size : status: 6 
NvMMLiteOpen : Block : BlockType = 261 
NvMMLiteBlockCreate : Block : BlockType = 261 
NvMMLiteBlockCreate : Block : BlockType = 261 
NvMMLiteBlockCreate : Block : BlockType = 261 
NVMEDIA: Reading vendor.tegra.display-size : status: 6 
NVMEDIA: Reading vendor.tegra.display-size : status: 6 
NvMMLiteBlockCreate : Block : BlockType = 261 
NvMMLiteBlockCreate : Block : BlockType = 261 
NVMEDIA: Reading vendor.tegra.display-size : status: 6 
NVMEDIA: Reading vendor.tegra.display-size : status: 6 
NVMEDIA: Reading vendor.tegra.display-size : status: 6 
NvMMLiteBlockCreate : Block : BlockType = 261 
NvMMLiteBlockCreate : Block : BlockType = 261 
NvMMLiteBlockCreate : Block : BlockType = 261 
Creating LL OSD context new
0:00:08.213980348 19755      0xd88ef70 ERROR                nvinfer gstnvinfer.cpp:511:gst_nvinfer_logger:<primary_gie_classifier> NvDsInferContext[UID 1]:parseBoundingBox(): Could not find output coverage layer for parsing objects
0:00:08.214094627 19755      0xd88ef70 ERROR                nvinfer gstnvinfer.cpp:511:gst_nvinfer_logger:<primary_gie_classifier> NvDsInferContext[UID 1]:fillDetectionOutput(): Failed to parse bboxes
0:00:08.214153575 19755      0xd88ef70 ERROR                nvinfer gstnvinfer.cpp:511:gst_nvinfer_logger:<primary_gie_classifier> NvDsInferContext[UID 1]:parseBoundingBox(): Could not find output coverage layer for parsing objects
0:00:08.214194025 19755      0xd88ef70 ERROR                nvinfer gstnvinfer.cpp:511:gst_nvinfer_logger:<primary_gie_classifier> NvDsInferContext[UID 1]:fillDetectionOutput(): Failed to parse bboxes
Segmentation fault (core dumped)

My errors: parseBoundingBox(): Could not find output coverage layer for parsing objects, fillDetectionOutput(): Failed to parse bboxes, nvinfer gstnvinfer.cpp:511:gst_nvinfer_logger:<primary_gie_classifier> NvDsInferContext[UID 1]:fillDetectionOutput(): Failed to parse bboxes
Segmentation fault (core dumped)