Hi, I am trying to run custom trained ssd_inception_2018 but currently the config file present in the repo is for ssd_inception_2017.
Does anyone know how we can obtain that config file or what changes should I do to make inception_2018 work?
Thanks

Hi,

what is required to get the ssd_mobilenet_v2_oid_v4_2018_12_12 model working?
Would it be possible to add a config for this model?

http://download.tensorflow.org/models/object_detection/ssd_mobilenet_v2_oid_v4_2018_12_12.tar.gz

While running ssd inception v2 coco model over a video, after running for 2 mins the process got stopped with Segmentation fault (core dumped) error on Jetson Nano

[TensorRT] INFO: Glob Size is 50485364 bytes.
[TensorRT] INFO: Added linear block of size 4292096
[TensorRT] INFO: Added linear block of size 1522176
[TensorRT] INFO: Added linear block of size 739328
[TensorRT] INFO: Added linear block of size 411648
[TensorRT] INFO: Added linear block of size 204800
[TensorRT] INFO: Added linear block of size 57344
[TensorRT] INFO: Added linear block of size 30720
[TensorRT] INFO: Added linear block of size 20992
[TensorRT] INFO: Added linear block of size 9728
[TensorRT] INFO: Added linear block of size 9216
[TensorRT] INFO: Added linear block of size 2560
[TensorRT] INFO: Added linear block of size 2560
[TensorRT] INFO: Added linear block of size 1024
[TensorRT] INFO: Added linear block of size 512
[TensorRT] INFO: Found Creator FlattenConcat_TRT
[TensorRT] INFO: Found Creator GridAnchor_TRT
[TensorRT] INFO: Found Creator FlattenConcat_TRT
[TensorRT] INFO: Found Creator NMS_TRT
[TensorRT] INFO: Deserialize required 3084695 microseconds.
Gtk-Message: 11:48:55.371: Failed to load module "canberra-gtk-module"
Segmentation fault (core dumped)

Hi
'''
namespace_plugin_map = {
"MultipleGridAnchorGenerator": PriorBox,
"Postprocessor": NMS,
"Preprocessor": Input,
"ToFloat": Input,
"image_tensor": Input,
"Concatenate": concat_priorbox,
"concat": concat_box_loc,
"concat_1": concat_box_conf
}

NMS is the corresponding namespace Postprocessor, but under the Postprocessor, there are other operations, such as the Preprocessor/map/while/Enter, unified into NMS.
Will this lead to problems in node connection？

When I compiled 'libflattenconcat.so' on ubuntu, it reported an error
I follow the process below：
$ {tensorrt}/samples/python/uff_ssd
$ mkdir build
$ cd build
$ cmake ../
$ make
I also try sudo to build it ,it still report errors

(base) wyl@wyl-Inspiron-3670:~/software/pc/TensorRT-6.0.1.5/samples/python/uff_ssd/build$ sudo cmake ../
-- The CXX compiler identification is GNU 7.4.0
-- Check for working CXX compiler: /usr/bin/c++
-- Check for working CXX compiler: /usr/bin/c++ -- works
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - done
-- Detecting CXX compile features
-- Detecting CXX compile features - done
-- Looking for C++ include pthread.h
-- Looking for C++ include pthread.h - found
-- Looking for pthread_create
-- Looking for pthread_create - not found
-- Looking for pthread_create in pthreads
-- Looking for pthread_create in pthreads - not found
-- Looking for pthread_create in pthread
-- Looking for pthread_create in pthread - found
-- Found Threads: TRUE
-- Found CUDA: /usr/local/cuda (found version "10.1")
-- Configurable variable TRT_LIB set to /usr/lib/x86_64-linux-gnu
-- Configurable variable TRT_INCLUDE set to /usr/include/x86_64-linux-gnu
-- Configurable variable CUDA_ROOT set to /usr/local/cuda

The following variables are derived from the values of the previous variables unless provided explicitly:

-- Configurable variable NVINFER_LIB set to _NVINFER_LIB-NOTFOUND
-- Configurable variable CUBLAS_LIB set to /usr/lib/x86_64-linux-gnu/libcublas.so
-- Configurable variable CUDA_INC_DIR set to /usr/local/cuda/include
CMake Error: The following variables are used in this project, but they are set to NOTFOUND.
Please set them or make sure they are set and tested correctly in the CMake files:
_NVINFER_LIB
linked by target "flattenconcat" in directory /home/wyl/software/pc/TensorRT-6.0.1.5/samples/python/uff_ssd

-- Configuring incomplete, errors occurred!
See also "/home/wyl/software/pc/TensorRT-6.0.1.5/samples/python/uff_ssd/build/CMakeFiles/CMakeOutput.log".

@AastaNV ,

while running the "python3 main.py object_7.jpg "
I am facing the below error,

/home/deeplearning/.local/lib/python3.6/site-packages/tensorflow/python/framework/dtypes.py:526: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.
_np_qint8 = np.dtype([("qint8", np.int8, 1)])
/home/deeplearning/.local/lib/python3.6/site-packages/tensorflow/python/framework/dtypes.py:527: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.
_np_quint8 = np.dtype([("quint8", np.uint8, 1)])
/home/deeplearning/.local/lib/python3.6/site-packages/tensorflow/python/framework/dtypes.py:528: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.
_np_qint16 = np.dtype([("qint16", np.int16, 1)])
/home/deeplearning/.local/lib/python3.6/site-packages/tensorflow/python/framework/dtypes.py:529: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.
_np_quint16 = np.dtype([("quint16", np.uint16, 1)])
/home/deeplearning/.local/lib/python3.6/site-packages/tensorflow/python/framework/dtypes.py:530: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.
_np_qint32 = np.dtype([("qint32", np.int32, 1)])
/home/deeplearning/.local/lib/python3.6/site-packages/tensorflow/python/framework/dtypes.py:535: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.
np_resource = np.dtype([("resource", np.ubyte, 1)])
Traceback (most recent call last):
File "main.py", line 18, in
ctypes.CDLL("/home/deeplearning/Downloads/TRT_object_detection/lib/libflattenconcat.so")
File "/usr/lib/python3.6/ctypes/init.py", line 348, in init
self._handle = _dlopen(self._name, mode)
OSError: libnvinfer.so.5: cannot open shared object file: No such file or directory

please advise

Hi @AastaNV

I got the following error:
OSError: lib/libflattenconcat.so: cannot open shared object file: No such file or directory

Pardon if this is a dumb question, but how is the create_node function supposed to look after the edits?

Mine looks like this:

def create_node(name, op=None, trt_plugin=False, **kwargs):
if not trt_plugin:
print("WARNING: To create TensorRT plugin nodes, please use the create_plugin_node function instead.")

    node = tf.NodeDef()
    node.name = name
    node.op = op if op else name
    node.attr["dtype"].type = 1

    for key, val in kwargs.items():
            if key == "dtype":
                    node.attr["dtype"].type = val.as_datatype_enum

    return update_node(node, name, op, trt_plugin, **kwargs)

But when I run the main.py code, it runs for a bit then it gives this error:

$ python3 main.py [image]
OpenCV Error: Assertion failed (scn == 3 || scn == 4) in cvtColor, file /home/nvidia/build_opencv/opencv/modules/imgproc/src/color.cpp, line 11016
Traceback (most recent call last):
File "main.py", line 81, in
image = cv2.cvtColor(ori, cv2.COLOR_BGR2RGB)
cv2.error: /home/nvidia/build_opencv/opencv/modules/imgproc/src/color.cpp:11016: error: (-215) scn == 3 || scn == 4 in function cvtColor

and the code block in question is:

ori = cv2.imread(sys.argv[1])
image = cv2.cvtColor(ori, cv2.COLOR_BGR2RGB) # this line right here
image = cv2.resize(image, (model.dims[2],model.dims[1]))

Hello,
I have built TRT_ssd_mobilenet_v2_coco_2018_03_29.bin successfully. However, when I use this file in the deepstream4.0, the error occurs as follows:

deepstream-app: nvdsiplugin_ssd.cpp:72: FlattenConcat::FlattenConcat(const void*, size_t): Assertion `mConcatAxisID == 1 || mConcatAxisID == 2 || mConcatAxisID == 3' failed.
Aborted (core dumped)

I tried to convert a SSD mobilenet v1 and got the following error

Using output node NMS
Converting to UFF graph
Warning: No conversion function registered for layer: NMS_TRT yet.
Converting NMS as custom op: NMS_TRT
Warning: No conversion function registered for layer: GridAnchor_TRT yet.
Converting GridAnchor as custom op: GridAnchor_TRT
Warning: No conversion function registered for layer: FlattenConcat_TRT yet.
Converting concat_box_loc as custom op: FlattenConcat_TRT
Traceback (most recent call last):
File "main.py", line 32, in
uff_model = uff.from_tensorflow(dynamic_graph.as_graph_def(), model.output_name, output_filename='tmp.uff')
File "/home/tonto/.virtualenvs/deep_learning/lib/python3.5/site-packages/uff/converters/tensorflow/conversion_helpers.py", line 181, in from_tensorflow
debug_mode=debug_mode)
File "/home/tonto/.virtualenvs/deep_learning/lib/python3.5/site-packages/uff/converters/tensorflow/converter.py", line 94, in convert_tf2uff_graph
uff_graph, input_replacements, debug_mode=debug_mode)
File "/home/tonto/.virtualenvs/deep_learning/lib/python3.5/site-packages/uff/converters/tensorflow/converter.py", line 79, in convert_tf2uff_node
op, name, tf_node, inputs, uff_graph, tf_nodes=tf_nodes, debug_mode=debug_mode)
File "/home/tonto/.virtualenvs/deep_learning/lib/python3.5/site-packages/uff/converters/tensorflow/converter.py", line 47, in convert_layer
return cls.registry_[op](name, tf_node, inputs, uff_graph, **kwargs)
File "/home/tonto/.virtualenvs/deep_learning/lib/python3.5/site-packages/uff/converters/tensorflow/converter_functions.py", line 21, in convert_placeholder
dtype = tf2uff.convert_tf2numpy_dtype(tf_node.attr['dtype'].type)
File "/home/tonto/.virtualenvs/deep_learning/lib/python3.5/site-packages/uff/converters/tensorflow/converter.py", line 103, in convert_tf2numpy_dtype
return tf.as_dtype(dtype).as_numpy_dtype
File "/home/tonto/.virtualenvs/deep_learning/lib/python3.5/site-packages/tensorflow/python/framework/dtypes.py", line 719, in as_dtype
raise TypeError("Cannot convert value %r to a TensorFlow DType." % type_value)
TypeError: Cannot convert value 0 to a TensorFlow DType.

does anyone know what's the issue?

I am using TF-GPU 1.13.1

Hi

I was able to create the libflattenconcat.so so thanks for that.

Now while running the code with an own model i get the error msg:

TypeError: Cannot convert value 0 to a TensorFlow DType.

Any ideas?

Thank you
Cheers Dario

Hi :

tensorRT :6.0.1.5
tensorflow:1.14
system:ubuntu 18.04
hardware：PC，NVIDIA 1060 3GB

1. When I run the original program, it reports an error
   'lib/libflattenconcat.so: cannot open shared object file: No such file or directory'
   so I changed the program to be：

from ctypes.util import find_library
libname = find_library("lib/libflattenconcat.so")
ctypes.CDLL(libname)

It still reports an error：

/home/wyl/.conda/envs/pytensorrt/bin/python /home/wyl/project/tensorrt/object_google_model/TRT_object_detection-master/main.py
/home/wyl/.conda/envs/pytensorrt/lib/python3.6/site-packages/tensorflow/python/framework/dtypes.py:516: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.
_np_qint8 = np.dtype([("qint8", np.int8, 1)])
/home/wyl/.conda/envs/pytensorrt/lib/python3.6/site-packages/tensorflow/python/framework/dtypes.py:517: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.
_np_quint8 = np.dtype([("quint8", np.uint8, 1)])
/home/wyl/.conda/envs/pytensorrt/lib/python3.6/site-packages/tensorflow/python/framework/dtypes.py:518: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.
_np_qint16 = np.dtype([("qint16", np.int16, 1)])
/home/wyl/.conda/envs/pytensorrt/lib/python3.6/site-packages/tensorflow/python/framework/dtypes.py:519: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.
_np_quint16 = np.dtype([("quint16", np.uint16, 1)])
/home/wyl/.conda/envs/pytensorrt/lib/python3.6/site-packages/tensorflow/python/framework/dtypes.py:520: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.
_np_qint32 = np.dtype([("qint32", np.int32, 1)])
/home/wyl/.conda/envs/pytensorrt/lib/python3.6/site-packages/tensorflow/python/framework/dtypes.py:525: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.
np_resource = np.dtype([("resource", np.ubyte, 1)])
/home/wyl/.conda/envs/pytensorrt/lib/python3.6/site-packages/tensorboard/compat/tensorflow_stub/dtypes.py:541: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.
_np_qint8 = np.dtype([("qint8", np.int8, 1)])
/home/wyl/.conda/envs/pytensorrt/lib/python3.6/site-packages/tensorboard/compat/tensorflow_stub/dtypes.py:542: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.
_np_quint8 = np.dtype([("quint8", np.uint8, 1)])
/home/wyl/.conda/envs/pytensorrt/lib/python3.6/site-packages/tensorboard/compat/tensorflow_stub/dtypes.py:543: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.
_np_qint16 = np.dtype([("qint16", np.int16, 1)])
/home/wyl/.conda/envs/pytensorrt/lib/python3.6/site-packages/tensorboard/compat/tensorflow_stub/dtypes.py:544: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.
_np_quint16 = np.dtype([("quint16", np.uint16, 1)])
/home/wyl/.conda/envs/pytensorrt/lib/python3.6/site-packages/tensorboard/compat/tensorflow_stub/dtypes.py:545: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.
_np_qint32 = np.dtype([("qint32", np.int32, 1)])
/home/wyl/.conda/envs/pytensorrt/lib/python3.6/site-packages/tensorboard/compat/tensorflow_stub/dtypes.py:550: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.
np_resource = np.dtype([("resource", np.ubyte, 1)])
WARNING:tensorflow:From /home/wyl/.conda/envs/pytensorrt/lib/python3.6/site-packages/uff/converters/tensorflow/conversion_helpers.py:18: The name tf.GraphDef is deprecated. Please use tf.compat.v1.GraphDef instead.

WARNING:tensorflow:From /home/wyl/.conda/envs/pytensorrt/lib/python3.6/site-packages/graphsurgeon/_utils.py:2: The name tf.NodeDef is deprecated. Please use tf.compat.v1.NodeDef instead.

WARNING:tensorflow:From /home/wyl/.conda/envs/pytensorrt/lib/python3.6/site-packages/graphsurgeon/StaticGraph.py:125: The name tf.gfile.GFile is deprecated. Please use tf.io.gfile.GFile instead.

WARNING: To create TensorRT plugin nodes, please use the create_plugin_node function instead.
UFF Version 0.6.5
=== Automatically deduced input nodes ===
[name: "Input"
op: "Placeholder"
attr {
key: "dtype"
value {
type: DT_FLOAT
}
}
attr {
key: "shape"
value {
shape {
dim {
size: 1
}
dim {
size: 3
}
dim {
size: 300
}
dim {
size: 300
}
}
}
}
]

Using output node NMS
Converting to UFF graph
Warning: No conversion function registered for layer: NMS_TRT yet.
Converting NMS as custom op: NMS_TRT
WARNING:tensorflow:From /home/wyl/.conda/envs/pytensorrt/lib/python3.6/site-packages/uff/converters/tensorflow/converter.py:179: The name tf.AttrValue is deprecated. Please use tf.compat.v1.AttrValue instead.

Warning: No conversion function registered for layer: GridAnchor_TRT yet.
Converting GridAnchor as custom op: GridAnchor_TRT
Warning: No conversion function registered for layer: FlattenConcat_TRT yet.
Converting concat_box_loc as custom op: FlattenConcat_TRT
Warning: No conversion function registered for layer: FlattenConcat_TRT yet.
Converting concat_box_conf as custom op: FlattenConcat_TRT
DEBUG [/home/wyl/.conda/envs/pytensorrt/lib/python3.6/site-packages/uff/converters/tensorflow/converter.py:96] Marking ['NMS'] as outputs
No. nodes: 1094
UFF Output written to tmp.uff
[TensorRT] INTERNAL ERROR: Assertion failed: mConcatAxisID == 1 || mConcatAxisID == 2 || mConcatAxisID == 3
flattenConcat.cpp:29
Aborting...

Process finished with exit code 134 (interrupted by signal 6: SIGABRT)

Hi, is it possible to add the faster_rcnn_resnet50_lowproposals_coco_2018_01_28 model ?

How can i do this?

Thanks,

Hi @AastaNV ,

After re-training the mobilenet with a different dataset (i'm using egohands for hands detection) and reducing the number of classes from 90 to 1, I'm getting the following error

NOTE: UFF has been tested with TensorFlow 1.12.0. Other versions are not guaranteed to work
WARNING: The version of TensorFlow installed on this system is not guaranteed to work with UFF.
UFF Version 0.6.3
=== Automatically deduced input nodes ===
[name: "Input"
op: "Placeholder"
input: "Cast"
attr {
  key: "dtype"
  value {
    type: DT_FLOAT
  }
}
attr {
  key: "shape"
  value {
    shape {
      dim {
        size: 1
      }
      dim {
        size: 3
      }
      dim {
        size: 300
      }
      dim {
        size: 300
      }
    }
  }
}
]
=========================================

Using output node NMS
Converting to UFF graph
Warning: No conversion function registered for layer: NMS_TRT yet.
Converting NMS as custom op: NMS_TRT
WARNING:tensorflow:From /usr/lib/python3.6/dist-packages/uff/converters/tensorflow/converter.py:179: The name tf.AttrValue is deprecated. Please use tf.compat.v1.AttrValue instead.

Warning: No conversion function registered for layer: FlattenConcat_TRT yet.
Converting concat_box_conf as custom op: FlattenConcat_TRT
Warning: No conversion function registered for layer: Cast yet.
Converting Cast as custom op: Cast
Warning: No conversion function registered for layer: GridAnchor_TRT yet.
Converting GridAnchor as custom op: GridAnchor_TRT
Warning: No conversion function registered for layer: FlattenConcat_TRT yet.
Converting concat_box_loc as custom op: FlattenConcat_TRT
No. nodes: 451
UFF Output written to tmp.uff
[TensorRT] ERROR: UffParser: Validator error: Cast: Unsupported operation _Cast
[TensorRT] ERROR: Network must have at least one output
Traceback (most recent call last):
  File "main.py", line 44, in <module>
    buf = engine.serialize()
AttributeError: 'NoneType' object has no attribute 'serialize'

Comparing to the default model, the Placeholder layer has an additional "Cast" input layer, which is not supported by TensorRT. I'm trying to figure out how to use the graphsurgeon to handle this. Could you please propose any suggestions?

I tried to use some other models not included in the list and run into an issue with the reshape operator producing an Volume Mismatch error.

Looking at this thread this seems to be a known issue with TensorRT not supporting this operator at this point. Is there a way to go around this with a plugging of some sort? Which models will you be supporting in the near future?

Thanks.

I could optimized model 'ssd_inception_v2_coco_2017_11_17' using main.py script and saved the engine file for TensorRT inference.

Now, for using this model in deepstream test app, I am providing the engine file in parameter file of the application as 'model-engine-file' along with 'libflattenconcat.so' as 'custom-lib-path'.

Issue is : Application starts and runs the video on sink but detection boxes are not there even if there are so many cars in the sample video provided in DeepStream SDK. One error is continously being printed - 'Error: Could not find coverage layer while parsing output'.

By default, deepstream only supports resnet parsers, so If I use SSD then I think I have to provide also function name in 'parse-bbox-func-name' if I use custom plugin for error but I have no source of this libflattenconcat.so, can you please provide the function name so that application can parse the outputs.

My System

Trace of cat /etc/nv_tegra_release

# R32 (release), REVISION: 4.2, GCID: 20074772, BOARD: t210ref, EABI: aarch64, DATE: Thu Apr 9 01:22:12 UTC 2020

What I've done?

1. I've downloaded the latest Jetpack SDk from here.

2. The file name was nv-jetson-nano-sd-card-image-r32.4.2.zip. Flashed it using Baleno Etcher.

3. Installed TensorFlow 1.13.1 using the command mentioned in this repository.

pip3 install --extra-index-url https://developer.download.nvidia.com/compute/redist/jp/v42 tensorflow-gpu==1.13.1+nv19.5 --user

4. Then comes the following error. TensorFlow 1.13.1 requires CUDA 10.0 and it can't find the same in the Jetson Nano.

ImportError: libcublas.so.10.0: cannot open shared object file: No such file or directory

5. Because, by default, the CUDA version installed in Jetson Nano is 10.2.

6. Another issue is TensorRT version. TensorFlow 1.13.1 is built for TensorRT 5.0. The default TensorRT version is 7.1.

Questions

1. Can I downgrade the CUDA version to 10.0 ? How to do that?

   • The CUDA version is not recommended to be downgraded.

2. Can I downgrade the TensorRT version to 5? How to do that?

   • There is a work around mentioned by @AastaNV on discussion forum which points out to this repository. What exactly it is?

3. How to make this repository work on r32.4.2?

Strangely, I am having this error when I try to run main.py on the Jetson-Nano. It main.py runs just fine on the PC.

Any suggestions would be very helpful. I am completely stuck because of this.!!!!

Can I ask how did you get/generate your "libflattenconcat.so" ?

I want to use this repo to detect object from video stream or from CSI camera. But I'm not sure whether I can and what to do to perform that idea. Can anyone share with me?

Hi,
Thank you for sharing the codes.
I used jetpack 4.2.2 and opencv 3.4.6 and tensorflow 1.12.2.
and follow your codes for converting the model to uff and tensorrt model. but the structural of your code don't support batch of image processing at the same time. if prediction time of one image takes 50 ms, the prediction time of batch of three image take 150 ms, why?
How do I solve the problem?

I get an error while installing tensorflow on jetson nano. here is the complete output

$ pip3 install --extra-index-url https://developer.download.nvidia.com/compute/redist/jp/v42 tensorflow-gpu==1.13.1+nv19.5 --user

Collecting tensorflow-gpu==1.13.1+nv19.5
Downloading https://developer.download.nvidia.com/compute/redist/jp/v42/tensorflow-gpu/tensorflow_gpu-1.13.1+nv19.5-cp36-cp36m-linux_aarch64.whl (204.6MB)
100% |████████████████████████████████| 204.6MB 2.5kB/s
Collecting numpy>=1.13.3 (from tensorflow-gpu==1.13.1+nv19.5)
Exception:
Traceback (most recent call last):
File "/usr/lib/python3/dist-packages/pip/basecommand.py", line 215, in main
status = self.run(options, args)
File "/usr/lib/python3/dist-packages/pip/commands/install.py", line 353, in run
wb.build(autobuilding=True)
File "/usr/lib/python3/dist-packages/pip/wheel.py", line 749, in build
self.requirement_set.prepare_files(self.finder)
File "/usr/lib/python3/dist-packages/pip/req/req_set.py", line 380, in prepare_files
ignore_dependencies=self.ignore_dependencies))
File "/usr/lib/python3/dist-packages/pip/req/req_set.py", line 554, in _prepare_file
require_hashes
File "/usr/lib/python3/dist-packages/pip/req/req_install.py", line 278, in populate_link
self.link = finder.find_requirement(self, upgrade)
File "/usr/lib/python3/dist-packages/pip/index.py", line 465, in find_requirement
all_candidates = self.find_all_candidates(req.name)
File "/usr/lib/python3/dist-packages/pip/index.py", line 423, in find_all_candidates
for page in self._get_pages(url_locations, project_name):
File "/usr/lib/python3/dist-packages/pip/index.py", line 568, in _get_pages
page = self._get_page(location)
File "/usr/lib/python3/dist-packages/pip/index.py", line 683, in _get_page
return HTMLPage.get_page(link, session=self.session)
File "/usr/lib/python3/dist-packages/pip/index.py", line 795, in get_page
resp.raise_for_status()
File "/usr/share/python-wheels/requests-2.18.4-py2.py3-none-any.whl/requests/models.py", line 935, in raise_for_status
raise HTTPError(http_error_msg, response=self)
requests.exceptions.HTTPError: 404 Client Error: Not Found for url: https://developer.download.nvidia.com/compute/redist/jp/v42/numpy/

I'm using cable connection to the board, internet seems to be working fine.

how do I update graphsurgeon converter I am confused to put the code that has been provided

diff --git a/node_manipulation.py b/node_manipulation.py
index d2d012a..1ef30a0 100644
--- a/node_manipulation.py
+++ b/node_manipulation.py
@@ -30,6 +30,7 @@ def create_node(name, op=None, _do_suffix=False, **kwargs):
node = NodeDef()
node.name = name
node.op = op if op else name

• node.attr["dtype"].type = 1
  for key, val in kwargs.items():
  if key == "dtype":
  node.attr["dtype"].type = val.as_datatype_enu