Tensorrt: [08/18/2019-20:58:10] [E] [TRT] UffParser: Validator error: mrcnn_mask_deconv/add_1: Unsupported operation _AddV2 &&&& FAILED TensorRT.sample_maskrcnn # ./sample_uff_maskRCNN -d

[jinfagang]

[08/18/2019-20:58:10] [E] [TRT] UffParser: Validator error: mrcnn_mask_deconv/add_1: Unsupported operation _AddV2
&&&& FAILED TensorRT.sample_maskrcnn # ./sample_uff_maskRCNN -d ~/data

ubuntu16.04
TensorRT 6.x (build source from git branch release/6.0)
following tutorial converts matterport maskrcnn model successfully to uff, inference got this result.

[xiaoniu1235]

Running up, I got the error：
[08/19/2019-16:29:24] [I] Building and running a GPU inference engine for Mask RCNN
[08/19/2019-16:29:34] [E] [TRT] UffParser: Validator error: fpn_p3upsampled: Unsupported operation _ResizeNearest_TRT

windows 10
TensorRT 6.x zip install

[jinfagang]

Well, why so many errors

[jinfagang]

先别想着训练了，能跑通实例就不错了。。。。。

don't think about training.... thanks to God if they can even run....

[Tyler-D]

Running up, I got the error：
[08/19/2019-16:29:24] [I] Building and running a GPU inference engine for Mask RCNN
[08/19/2019-16:29:34] [E] [TRT] UffParser: Validator error: fpn_p3upsampled: Unsupported operation _ResizeNearest_TRT

windows 10
TensorRT 6.x zip install

Hi @xiaoniu1235 , Can you check if your excution program linked to the correct plugin library ?

[Tyler-D]

[08/18/2019-20:58:10] [E] [TRT] UffParser: Validator error: mrcnn_mask_deconv/add_1: Unsupported operation _AddV2
&&&& FAILED TensorRT.sample_maskrcnn # ./sample_uff_maskRCNN -d ~/data

ubuntu16.04
TensorRT 6.x (build source from git branch release/6.0)
following tutorial converts matterport maskrcnn model successfully to uff, inference got this result.

@jinfagang What's the version of your tensorflow when you do the conversion?

[xiaoniu1235]

Running up, I got the error：
[08/19/2019-16:29:24] [I] Building and running a GPU inference engine for Mask RCNN
[08/19/2019-16:29:34] [E] [TRT] UffParser: Validator error: fpn_p3upsampled: Unsupported operation _ResizeNearest_TRT
windows 10
TensorRT 6.x zip install

Hi @xiaoniu1235 , Can you check if your excution program linked to the correct plugin library ?

Sorry,i can't generate a build system for CMake, so i use tensorrt zip! windows 10.
I copied the build runtime environment for sampleFasterRCNN, but I didn't see the use of plugin in sampleUffMaskRCNN.cpp and mrcnn_config.h.

[Tyler-D]

@xiaoniu1235 , So that's the point. UffMaskRCNN uses some plugin layers provided in this repo including this _ResizeNearest_TRT in your error msg.

[jinfagang]

What does this error mean: mrcnn_mask_deconv/add_1: Unsupported operation _AddV2?

[xiaoniu1235]

@Tyler-D
I tracked the code and the parse mrcnn_nchw.uff file went wrong!
Parser is created using nvuffparser::createUffParser(), it seems that there is no plugin, compare the sampleFasterRCNN.cpp code

sampleUffMaskRCNN.cpp
auto parser = SampleUniquePtr(nvuffparser::createUffParser());
if (!parser)
{
return false;
}

auto constructed = constructNetwork(builder, network, parser);
if (!constructed)
{
    return false;
}
...

sampleFasterRCNN.cpp
auto parser = SampleUniquePtr(nvcaffeparser1::createCaffeParser());
if (!parser)
{
return false;
}

FRCNNPluginFactory pluginFactory;
parser->setPluginFactoryV2(&pluginFactory);
constructNetwork(parser, builder, network, config);

sampleFasterRCNN can clearly see the call of plugin, but sampleUffMaskRCNN I did not find the corresponding code, is I wrong?

[Tyler-D]

@jinfagang No operation called "AddV2" supported in TensorRT. https://docs.nvidia.com/deeplearning/sdk/tensorrt-support-matrix/index.html You can check the supported layers here. I noticed tensoflow has updated to 1.15 recently. So it is possible it adds some computation nodes that TensorRT uff cannot handle right now. You may downgrade to 1.14 and I guess this node will become "Add" and compatible with UFF parser.

[Tyler-D]

@xiaoniu1235 The schema to use plugin layer has changed a little with version updating. Now we use https://github.com/NVIDIA/TensorRT/blob/07ed9b57b1ff7c24664388e5564b17f7ce2873e5/samples/opensource/sampleUffMaskRCNN/sampleUffMaskRCNN.cpp#L352 instead of plugin factory to initialize plugin layers.

[jinfagang]

@Tyler-D I installed tensorflow 1.14 build from source. And I traced uff model succesffuly.
How Nvidia can traced a model with unsupported node name? Haven't they check available nodes like ONNX converts does? I don't think this is an tensorflow version issue here. Maybe it's uff version issue? I simply install uff from TensorRT6.0 tar file.

[guods]

@xiaoniu1235 Did you build the TensorRT6 succesffuly ? I meet the problems when I build, such as RT_LIB_NOTFOUND and uff , I don't know what is the RT_LIB; Nvidia don't open uff code, do you know the complied binaries about UFF ?

[dssdyx]

@Tyler-D I installed tensorflow 1.14 build from source. And I traced uff model succesffuly.
How Nvidia can traced a model with unsupported node name? Haven't they check available nodes like ONNX converts does? I don't think this is an tensorflow version issue here. Maybe it's uff version issue? I simply install uff from TensorRT6.0 tar file.

Have you solved this problem?

[dssdyx]

@jinfagang No operation called "AddV2" supported in TensorRT. https://docs.nvidia.com/deeplearning/sdk/tensorrt-support-matrix/index.html You can check the supported layers here. I noticed tensoflow has updated to 1.15 recently. So it is possible it adds some computation nodes that TensorRT uff cannot handle right now. You may downgrade to 1.14 and I guess this node will become "Add" and compatible with UFF parser.

Maybe you are right. I use tf1.15.0-rc2 there are some ops changed from Add to AddV2

[dssdyx]

@Tyler-D I installed tensorflow 1.14 build from source. And I traced uff model succesffuly.
How Nvidia can traced a model with unsupported node name? Haven't they check available nodes like ONNX converts does? I don't think this is an tensorflow version issue here. Maybe it's uff version issue? I simply install uff from TensorRT6.0 tar file.

I use tf1.9 now and the problem with AddV2 disappears, so I think it's version conflict. But now when I run sample_uff_maskRCNN, it shows that "[E] [TRT] UffParser: Validator error: mrcnn_detection_bboxes: Unsupported operation _SpecialSlice_TRT" which I think it's because that the plugins were not correctly installed. I plan to rewrite relative plugins like these are released in TensorRT6.0.1.

[jinfagang]

@dssdyx Or maybe tf1.9 to low

[dssdyx]

@jinfagang Yes, you can also try other versions. And do not use the latest version 1.15.0-rc2 and 2.0.0.

[dssdyx]

@jinfagang the performance is 350ms/frame in 1050Ti

[jinfagang]

@dssdyx That is too slow. 2.8 fps..

Can u have a test on GTX1080ti or RTX2080ti? Would that able up to 20 fps?

[dssdyx]

@jinfagang I don't have better device.

[guods]

“mrcnn_mask_deconv/add_1: Unsupported operation _AddV2”，did you solved the preblem?

[guods]

“mrcnn_mask_deconv/add_1: Unsupported operation _AddV2”，did you solved the preblem?

[dssdyx]

“mrcnn_mask_deconv/add_1: Unsupported operation _AddV2”，did you solved the preblem?

Install lower verison tf such as tf1.9

[jinfagang]

@dssdyx Would u like share a uff to test on more decent gpus for the performance checking? I am currently only have tensorflow 2.0 with cuda10.1 installed.

[rtrobin]

@xiaoniu1235 , So that's the point. UffMaskRCNN uses some plugin layers provided in this repo including this _ResizeNearest_TRT in your error msg.

Hi @xiaoniu1235 . I'm facing a similar problem here. I found the plugin layers needed for MaskRCNN were provided in this repo, plugin sub-folder. Do you have any idea how to link the custom plugins in this repo to the MaskRCNN sample?

[dssdyx]

@dssdyx Would u like share a uff to test on more decent gpus for the performance checking? I am currently only have tensorflow 2.0 with cuda10.1 installed.

I don't think it will work in your env.

[gemmit]

@dssdyx Did it work for you only by reinstalling a different version of TensorFlow or did you do further changes?

[gemmit]

@dssdyx Did it work for you only by reinstalling a different version of TensorFlow or did you do further changes?

Works for me now. I am able to convert the model and also run the maskrcnn example.
Thanks

[rmccorm4]

Just sharing this here for visibility - @gemmit was able to run the sample with the following configuration, thanks for sharing :slightly_smiling_face: :

TensorRT: 6.0.1.5
CUDA: 9.0
CuDNN: 7.6.3
python: 3.6.5
Tensorflow-gpu==1.9.0
Keras: 2.1.3

I was able to convert the model using uff-converter 
and run the mask-rcnn example as mentioned here

Thanks

If anyone is able to run it successfully with a more recent version of tensorflow-gpu, please share your full environment configuration so others can get it working as well!

---

FYI it seems these issues stemmed from (1) pip install tensorflow-gpu not being compatible with CUDA 10.1 and (2) TF >= 1.15 turning add ops into addV2. See https://github.com/NVIDIA/TensorRT/issues/123 for more comprehensive details

[doomb007]

[08/18/2019-20:58:10] [E] [TRT] UffParser: Validator error: mrcnn_mask_deconv/add_1: Unsupported operation _AddV2
&&&& FAILED TensorRT.sample_maskrcnn # ./sample_uff_maskRCNN -d ~/data

ubuntu16.04
TensorRT 6.x (build source from git branch release/6.0)
following tutorial converts matterport maskrcnn model successfully to uff, inference got this result.

Hi @jinfagang, I have the same problem as you now.

If the command './sample_uff_maskRCNN' was used, the terminal display:
&&&& RUNNING TensorRT.sample_maskrcnn # ./sample_uff_maskRCNN
[10/12/2019-14:16:43] [I] Building and running a GPU inference engine for Mask RCNN
[10/12/2019-14:16:43] [E] [TRT] UffParser: Validator error: mrcnn_mask_deconv/add_1: Unsupported operation _AddV2
&&&& FAILED TensorRT.sample_maskrcnn # ./sample_uff_maskRCNN

And if the command 'sudo ./sample_uff_maskRCNN' was used, the terminal display:
./sample_uff_maskRCNN: error while loading shared libraries: libnvparsers.so.6: cannot open shared object file: No such file or directory

The strange thing is that: the scripts ./sample_googlenet and ./sample_uff_mnist can run successfully, which proves that my software environment should be suitable.

Is your problem solved? Can you share your experience? thank you.

[rmccorm4]

@doomb007

1. The error while loading shared libraries: libnvparsers.so.6: cannot open shared object file: No such file or directory error when running with sudo is somewhat expected, because when you run as sudo, your environment variables are not the same as your regular user. So things like LD_LIBRARY* won't be set correctly. I think that can be fixed with the -E or --preserve-env flag, although I don't think you need to be running the command with sudo.

2. I'm starting to get confused about that AddV2, error. I was messing around with this a week or two ago, and confirmed for myself that this only happened with tensorflow >= 1.15, though I was unable to find documentation supporting this. But I believe if you're using tensorflow <= 1.14, you shouldn't get this error. Is this not true for you? Can you share the output of pip freeze from the environment you're getting this error in?

[jinfagang]

@rmccorm4 Yeah, Nvidia guys written a bounch of steps but not specific exactly version of tensorflow they were using and assuming all tensorflow under 1.19 were same behavior, obviously this is not tensorflow style, it always changes a lot ....

[doomb007]

@rmccorm4

1. Got it!

2. Actually, I'm using Tensorflow-gpu1.14.0 at the moment(Before that, I used Tensorflow-gpu1.15.0 and Tensorflow-gpu2.0.0), don't know what I missed. the output of pip freeze is as follow:

absl-py==0.8.1
#alabaster==0.7.12
anaconda-client==1.7.2
anaconda-navigator==1.9.7
anaconda-project==0.8.3
asn1crypto==0.24.0
astor==0.8.0
astroid==2.2.5
astropy==3.2.1
atomicwrites==1.3.0
###attrs==19.1.0
Babel==2.7.0
backcall==0.1.0
backports.functools-lru-cache==1.5
backports.os==0.1.1
backports.shutil-get-terminal-size==1.0.0
backports.tempfile==1.0
backports.weakref==1.0.post1
beautifulsoup4==4.7.1
bitarray==0.9.3
bkcharts==0.2
bleach==3.1.0
bokeh==1.2.0
boto==2.49.0
Bottleneck==1.2.1
cachetools==3.1.1
certifi==2019.6.16
cffi==1.12.3
chardet==3.0.4
Click==7.0
cloudpickle==1.2.1
clyent==1.2.2
colorama==0.4.1
conda==4.7.12
conda-build==3.18.8
conda-package-handling==1.3.11
conda-verify==3.4.2
contextlib2==0.5.5
cryptography==2.7
cycler==0.10.0
Cython==0.29.12
cytoolz==0.10.0
dask==2.1.0
decorator==4.4.0
defusedxml==0.6.0
distributed==2.1.0
docutils==0.14
entrypoints==0.3
et-xmlfile==1.0.1
fastcache==1.1.0
filelock==3.0.12
Flask==1.1.1
future==0.17.1
gast==0.2.2
gevent==1.4.0
glob2==0.7
gmpy2==2.0.8
google-auth==1.6.3
google-auth-oauthlib==0.4.1
google-pasta==0.1.7
graphsurgeon==0.4.1
greenlet==0.4.15
grpcio==1.24.3
h5py==2.9.0
heapdict==1.0.0
html5lib==1.0.1
idna==2.8
imageio==2.5.0
imagesize==1.1.0
importlib-metadata==0.17
ipykernel==5.1.1
ipython==7.6.1
ipython-genutils==0.2.0
ipywidgets==7.5.0
isort==4.3.21
itsdangerous==1.1.0
jdcal==1.4.1
jedi==0.13.3
jeepney==0.4
Jinja2==2.10.1
joblib==0.13.2
json5==0.8.4
jsonschema==3.0.1
jupyter==1.0.0
jupyter-client==5.3.1
jupyter-console==6.0.0
jupyter-core==4.5.0
jupyterlab==1.0.2
jupyterlab-server==1.0.0
Keras==2.1.3
Keras-Applications==1.0.8
Keras-Preprocessing==1.1.0
keyring==18.0.0
kiwisolver==1.1.0
lazy-object-proxy==1.4.1
libarchive-c==2.8
lief==0.9.0
llvmlite==0.29.0
locket==0.2.0
lxml==4.3.4
Markdown==3.1.1
MarkupSafe==1.1.1
matplotlib==3.1.0
mccabe==0.6.1
mistune==0.8.4
mkl-fft==1.0.12
mkl-random==1.0.2
mkl-service==2.0.2
mock==3.0.5
more-itertools==7.0.0
mpmath==1.1.0
msgpack==0.6.1
multipledispatch==0.6.0
navigator-updater==0.2.1
nbconvert==5.5.0
nbformat==4.4.0
networkx==2.3
nltk==3.4.4
nose==1.3.7
notebook==6.0.0
numba==0.44.1
numexpr==2.6.9
numpy==1.16.4
numpydoc==0.9.1
oauthlib==3.1.0
olefile==0.46
openpyxl==2.6.2
opt-einsum==3.1.0
packaging==19.0
pandas==0.24.2
pandocfilters==1.4.2
parso==0.5.0
partd==1.0.0
path.py==12.0.1
pathlib2==2.3.4
patsy==0.5.1
pep8==1.7.1
pexpect==4.7.0
pickleshare==0.7.5
Pillow==6.1.0
pkginfo==1.5.0.1
pluggy==0.12.0
ply==3.11
prometheus-client==0.7.1
prompt-toolkit==2.0.9
protobuf==3.10.0
psutil==5.6.3
ptyprocess==0.6.0
py==1.8.0
pyasn1==0.4.7
pyasn1-modules==0.2.7
pycodestyle==2.5.0
pycosat==0.6.3
pycparser==2.19
pycrypto==2.6.1
pycurl==7.43.0.3
pyflakes==2.1.1
Pygments==2.4.2
pylint==2.3.1
pyodbc==4.0.26
pyOpenSSL==19.0.0
pyparsing==2.4.0
pyrsistent==0.14.11
PySocks==1.7.0
pytest==5.0.1
pytest-arraydiff==0.3
pytest-astropy==0.5.0
pytest-doctestplus==0.3.0
pytest-openfiles==0.3.2
pytest-remotedata==0.3.1
python-dateutil==2.8.0
pytz==2019.1
PyWavelets==1.0.3
PyYAML==5.1.1
pyzmq==18.0.0
QtAwesome==0.5.7
qtconsole==4.5.1
QtPy==1.8.0
requests==2.22.0
requests-oauthlib==1.2.0
rope==0.14.0
rsa==4.0
ruamel-yaml==0.15.46
scikit-image==0.15.0
scikit-learn==0.21.2
scipy==1.3.0
seaborn==0.9.0
SecretStorage==3.1.1
Send2Trash==1.5.0
simplegeneric==0.8.1
singledispatch==3.4.0.3
six==1.12.0
snowballstemmer==1.9.0
sortedcollections==1.1.2
sortedcontainers==2.1.0
soupsieve==1.8
Sphinx==2.1.2
sphinxcontrib-applehelp==1.0.1
sphinxcontrib-devhelp==1.0.1
sphinxcontrib-htmlhelp==1.0.2
sphinxcontrib-jsmath==1.0.1
sphinxcontrib-qthelp==1.0.2
sphinxcontrib-serializinghtml==1.1.3
sphinxcontrib-websupport==1.1.2
spyder==3.3.6
spyder-kernels==0.5.1
SQLAlchemy==1.3.5
statsmodels==0.10.0
sympy==1.4
tables==3.5.2
tblib==1.4.0
tensorboard==1.14.0
tensorflow-estimator==1.14.0
tensorflow-gpu==1.14.0
tensorrt==6.0.1.5
termcolor==1.1.0
terminado==0.8.2
testpath==0.4.2
toolz==0.10.0
tornado==6.0.3
tqdm==4.32.1
traitlets==4.3.2
uff==0.6.5
unicodecsv==0.14.1
urllib3==1.24.2
wcwidth==0.1.7
webencodings==0.5.1
Werkzeug==0.15.4
widgetsnbextension==3.5.0
wrapt==1.11.2
wurlitzer==1.0.2
xlrd==1.2.0
XlsxWriter==1.1.8
xlwt==1.3.0
zict==1.0.0
zipp==0.5.1

[doomb007]

@rmccorm4 @jinfagang

I think this error UffParser: Validator error: mrcnn_mask_deconv/add_1: Unsupported operation _AddV2 caused by wrong uff model mrcnn_nchw.uff. because my uff model was generated in the environment: TensorRT6.0+CUDA10.0+cuDNN7.6.3+Tensorflow1.15.0, It seems that the tensorflow1.15.0 is not suitable here.

After I regenerate the model mrcnn_nchw.uff use the script mrcnn_to_trt_single.py in the environment:TensorRT6.0+CUDA10.0+cuDNN7.6.3+Tensorflow1.14.0, and run ./sample_uff_maskRCNN, this error disappeared.

But another error has occurred: [10/13/2019-09:20:08] [E] [TRT] ../rtSafe/safeRuntime.cpp (25) - Cuda Error in allocate: 2 (out of memory).

[rmccorm4]

@doomb007 You can watch your memory usage in a separate shell with something like: watch -n 0.5 nvidia-smi (on Linux)

I'm pretty sure I ran this sample in my earlier posts on an 8GB P4 GPU, so I think 8GB should be enough memory for this. Make sure you don't have other processes using up too much GPU memory.

[jinfagang]

@doomb007 OK, seems tensorflow 1.14 needed here. You can have another test on a bigger memory GPU or shrink down batchsize, if you succeed, remember post some speed test (fps) here, I assume it can achieve 20 fps on GTX1080TI. If not, then too slow.

[doomb007]

@rmccorm4

I ran the sample on my personal computer, with a 2GB GPU(MX250), This may be the root of the error, I will try the sample on a workstation.

[rtrobin]

@doomb007 try fp16 mode.

[doomb007]

@jinfagang, OK

[doomb007]

@rtrobin thanks for your advice.

when run the command ./sample_uff_maskRCNN --fp16, the terminal display:

&&&& RUNNING TensorRT.sample_maskrcnn # ./sample_uff_maskRCNN --fp16
[10/13/2019-11:52:11] [I] Building and running a GPU inference engine for Mask RCNN
[10/13/2019-11:52:14] [W] [TRT] Half2 support requested on hardware without native FP16 support, performance will be negatively affected.
[10/13/2019-11:52:15] [E] [TRT] ../rtSafe/safeRuntime.cpp (25) - Cuda Error in allocate: 2 (out of memory)
[10/13/2019-11:52:15] [E] [TRT] ../rtSafe/safeRuntime.cpp (25) - Cuda Error in allocate: 2 (out of memory)
&&&& FAILED TensorRT.sample_maskrcnn # ./sample_uff_maskRCNN --fp16

[rmccorm4]

Hi @doomb007,

I'd be surprised if you'd be able to run any substantial workflow on a 2GB GPU, other than something like MNIST, even with --fp16. Please do let me know if you can run it on your workstation, and either way please share the GPU/Memory specs of your workstation as well after running it so that others have a reference for whether or not that configuration can handle the sample.

Also, FYI:

[10/13/2019-11:52:14] [W] [TRT] Half2 support requested on hardware without native FP16 support, performance will be negatively affected.

Only certain hardware platforms support FP16 acceleration, see the support matrix here: https://docs.nvidia.com/deeplearning/sdk/tensorrt-support-matrix/index.html#hardware-precision-matrix

[doomb007]

@rmccorm4

OK, Will share details after ran the sample on my workstation, Thank you!

[hwh-hit]

cuda=9.0  
cudnn = 7.6.4
keras=2.1.3
tensorflow-gpu=1.12

I can convert h5 to uff successfully, but when i run

make -j$(nproc)

in build folder,some error occurs:

TensorRT/samples/common/sampleDevice.h:79:60: error: ‘cudaLaunchHostFunc’ was not declared in this scope
         cudaCheck(cudaLaunchHostFunc(mStream, cudaSleep, ms));

but the error doen't matter maybe, because bin (sample_uff_maskRCNN) already generated successfully,

Speed:
1024x1024x3,1080gpu:98ms
512x512x3, 1080gpu : 46.1008ms

if someone can help me resolove the problem, that's would be great

[doomb007]

Hi @hwh-hit

I have not encountered a similar situation as you. You can even run the program, this error should not matter. And you'd better consult rmccorm4 if you want more professional advice.

[rmccorm4]

The above issue is being tracked here: https://github.com/NVIDIA/TensorRT/issues/221

[doomb007]

Hi @rmccorm4

I can run the MaskRCNN sample on a workstation with a 1080Ti GPU now. yes, the performance of 2G GPU is not enough to run the sample. I will test the speed of MaskRCNN inference next step.

[sugger23]

Hi @rmccorm4
I want to run this sample on windows10, I have got the h5touff file successfully, when I run the MaskRCNN sample ,some errors occurred,such as:
UffParser: Validator error: fpn_p3upsampled: Unsupported operation _ResizeNearest_TRT

then, I add the _ResizeNearest, SpecialSlice, PyramidROIAlign,ProposalLayer,DetectionLayer plugins,
the code is here:
initializePlugin(&gLogger.getTRTLogger(), "");
initializePlugin(&gLogger.getTRTLogger(), "");
initializePlugin(&gLogger.getTRTLogger(), "");
initializePlugin(&gLogger.getTRTLogger(), "");
initializePlugin(&gLogger.getTRTLogger(), "");

then I run the sample again, the errors "Unsupported operation" disappeared, bug, when ,it run function:parser->parse, the program Downtime，and no any prompt, can you give me some idea?
and can this sample run on window10?

[ApoorvaSuresh]

Hello all,

I successfully converted my model to UFF.

When I run the sample, I have this error:

&&&& RUNNING TensorRT.sample_maskrcnn # ./sample_maskRCNN -d /mnt/DTAA_data/DTAA/code/z662728/optimization/data/
[05/04/2020-09:24:19] [I] Building and running a GPU inference engine for Mask RCNN
[05/04/2020-09:24:21] [E] [TRT] UffParser: Validator error: mrcnn_detection_bboxes: Unsupported operation _SpecialSlice_TRT
&&&& FAILED TensorRT.sample_maskrcnn # ./sample_maskRCNN -d /mnt/DTAA_data/DTAA/code/z662728/optimization/data/

How did you solve it? @rmccorm4 @ShawnNew I would be very grateful if you could please help me :)

Config:
TF-gpu 1.14.0
CUDA 10.0
TensorRT: 6.0.1.5
Keras 2.1.3

Thanks!

[ApoorvaSuresh]

Hello all,

Another try with a different configuration:
I successfully converted my model to UFF. When I run the sample, I get the same error:

&&&& RUNNING TensorRT.sample_maskrcnn # ./sample_maskRCNN -d /mnt/DTAA_data/DTAA/code/z662728/optimization/data/
[05/04/2020-09:24:19] [I] Building and running a GPU inference engine for Mask RCNN
[05/04/2020-09:24:21] [E] [TRT] UffParser: Validator error: mrcnn_detection_bboxes: Unsupported operation _SpecialSlice_TRT
&&&& FAILED TensorRT.sample_maskrcnn # ./sample_maskRCNN -d /mnt/DTAA_data/DTAA/code/z662728/optimization/data/

How did you solve it? @gemmit I would be very grateful if you could please help me :)

Config:
TF-gpu 1.9
CUDA 9.0
TensorRT: 6.0.1.5
Keras 2.1.3
python 3.6.10

Thanks!

[qualia0000]

Running up, I got the error：
[08/19/2019-16:29:24] [I] Building and running a GPU inference engine for Mask RCNN
[08/19/2019-16:29:34] [E] [TRT] UffParser: Validator error: fpn_p3upsampled: Unsupported operation _ResizeNearest_TRT

windows 10
TensorRT 6.x zip install

Hello. May I know if you have solved the problem? I met the exact same error as yours under TRT6.x, Win10.