Converting tensorflow frozen rnn model to tensorRT graph with errors

Platform

Jetson AGX Xavier, with Jetpack 4.2.3 DP, python 3.6.9

cuda 10.0
cudnn 7.5
tensorrt 5.1.6.1
tensorflow-gpu 1.14

Problems

I would like to convert the frozen tensorflow pb model to tensorrt graph as instruction 2.2.2. TF-TRT 1.x Workflow With A Frozen Graph here -> https://docs.nvidia.com/deeplearning/frameworks/tf-trt-user-guide/index.html#using-savedmodel. I have already converted one cnn model and run it successfully. But, when I was trying to convert a rnn model including rnn.static_rnn layer, the following command would neither be completed nor report any errors.

trt_graph = converter.convert()

The process stoped but was not terminated as follows:

2019-12-04 14:32:27.916300: W tensorflow/core/platform/profile_utils/cpu_utils.cc:98] Failed to find bogomips in /proc/cpuinfo; cannot determine CPU frequency
2019-12-04 14:32:27.917870: I tensorflow/compiler/xla/service/service.cc:168] XLA service 0x3f461920 executing computations on platform Host. Devices:
2019-12-04 14:32:27.917998: I tensorflow/compiler/xla/service/service.cc:175]   StreamExecutor device (0): <undefined>, <undefined>
2019-12-04 14:32:27.927531: I tensorflow/stream_executor/platform/default/dso_loader.cc:42] Successfully opened dynamic library libcuda.so.1
2019-12-04 14:32:28.020732: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:972] ARM64 does not support NUMA - returning NUMA node zero
2019-12-04 14:32:28.021254: I tensorflow/compiler/xla/service/service.cc:168] XLA service 0x3f43b090 executing computations on platform CUDA. Devices:
2019-12-04 14:32:28.021478: I tensorflow/compiler/xla/service/service.cc:175]   StreamExecutor device (0): Xavier, Compute Capability 7.2
2019-12-04 14:32:28.022281: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:972] ARM64 does not support NUMA - returning NUMA node zero
2019-12-04 14:32:28.022478: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1640] Found device 0 with properties: 
name: Xavier major: 7 minor: 2 memoryClockRate(GHz): 1.5
pciBusID: 0000:00:00.0
2019-12-04 14:32:28.022582: I tensorflow/stream_executor/platform/default/dso_loader.cc:42] Successfully opened dynamic library libcudart.so.10.0
2019-12-04 14:32:28.022783: I tensorflow/stream_executor/platform/default/dso_loader.cc:42] Successfully opened dynamic library libcublas.so.10.0
2019-12-04 14:32:28.022894: I tensorflow/stream_executor/platform/default/dso_loader.cc:42] Successfully opened dynamic library libcufft.so.10.0
2019-12-04 14:32:28.022997: I tensorflow/stream_executor/platform/default/dso_loader.cc:42] Successfully opened dynamic library libcurand.so.10.0
2019-12-04 14:32:28.028037: I tensorflow/stream_executor/platform/default/dso_loader.cc:42] Successfully opened dynamic library libcusolver.so.10.0
2019-12-04 14:32:28.032048: I tensorflow/stream_executor/platform/default/dso_loader.cc:42] Successfully opened dynamic library libcusparse.so.10.0
2019-12-04 14:32:28.032289: I tensorflow/stream_executor/platform/default/dso_loader.cc:42] Successfully opened dynamic library libcudnn.so.7
2019-12-04 14:32:28.032619: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:972] ARM64 does not support NUMA - returning NUMA node zero
2019-12-04 14:32:28.032988: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:972] ARM64 does not support NUMA - returning NUMA node zero
2019-12-04 14:32:28.033167: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1763] Adding visible gpu devices: 0
2019-12-04 14:32:30.050452: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1181] Device interconnect StreamExecutor with strength 1 edge matrix:
2019-12-04 14:32:30.050661: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1187]      0 
2019-12-04 14:32:30.050723: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1200] 0:   N 
2019-12-04 14:32:30.051349: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:972] ARM64 does not support NUMA - returning NUMA node zero
2019-12-04 14:32:30.051807: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:972] ARM64 does not support NUMA - returning NUMA node zero
2019-12-04 14:32:30.052108: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1326] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 1255 MB memory) -> physical GPU (device: 0, name: Xavier, pci bus id: 0000:00:00.0, compute capability: 7.2)
WARNING:tensorflow:From save_model_ckpt.py:149: The name tf.GraphDef is deprecated. Please use tf.compat.v1.GraphDef instead.

2019-12-04 14:32:38.690353: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:972] ARM64 does not support NUMA - returning NUMA node zero
2019-12-04 14:32:38.690693: I tensorflow/core/grappler/devices.cc:55] Number of eligible GPUs (core count >= 8, compute capability >= 0.0): 1
2019-12-04 14:32:38.691224: I tensorflow/core/grappler/clusters/single_machine.cc:359] Starting new session
2019-12-04 14:32:38.693547: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:972] ARM64 does not support NUMA - returning NUMA node zero
2019-12-04 14:32:38.693818: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1640] Found device 0 with properties: 
name: Xavier major: 7 minor: 2 memoryClockRate(GHz): 1.5
pciBusID: 0000:00:00.0
2019-12-04 14:32:38.693978: I tensorflow/stream_executor/platform/default/dso_loader.cc:42] Successfully opened dynamic library libcudart.so.10.0
2019-12-04 14:32:38.694061: I tensorflow/stream_executor/platform/default/dso_loader.cc:42] Successfully opened dynamic library libcublas.so.10.0
2019-12-04 14:32:38.694130: I tensorflow/stream_executor/platform/default/dso_loader.cc:42] Successfully opened dynamic library libcufft.so.10.0
2019-12-04 14:32:38.694195: I tensorflow/stream_executor/platform/default/dso_loader.cc:42] Successfully opened dynamic library libcurand.so.10.0
2019-12-04 14:32:38.694268: I tensorflow/stream_executor/platform/default/dso_loader.cc:42] Successfully opened dynamic library libcusolver.so.10.0
2019-12-04 14:32:38.694332: I tensorflow/stream_executor/platform/default/dso_loader.cc:42] Successfully opened dynamic library libcusparse.so.10.0
2019-12-04 14:32:38.694393: I tensorflow/stream_executor/platform/default/dso_loader.cc:42] Successfully opened dynamic library libcudnn.so.7
2019-12-04 14:32:38.694577: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:972] ARM64 does not support NUMA - returning NUMA node zero
2019-12-04 14:32:38.694830: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:972] ARM64 does not support NUMA - returning NUMA node zero
2019-12-04 14:32:38.695052: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1763] Adding visible gpu devices: 0
2019-12-04 14:32:38.695320: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1181] Device interconnect StreamExecutor with strength 1 edge matrix:
2019-12-04 14:32:38.695420: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1187]      0 
2019-12-04 14:32:38.695499: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1200] 0:   N 
2019-12-04 14:32:38.695769: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:972] ARM64 does not support NUMA - returning NUMA node zero
2019-12-04 14:32:38.696102: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:972] ARM64 does not support NUMA - returning NUMA node zero
2019-12-04 14:32:38.696274: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1326] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 1255 MB memory) -> physical GPU (device: 0, name: Xavier, pci bus id: 0000:00:00.0, compute capability: 7.2)
2019-12-04 14:32:43.858438: I tensorflow/compiler/tf2tensorrt/segment/segment.cc:460] There are 7 ops of 6 different types in the graph that are not converted to TensorRT: ArgMax, Transpose, ConcatV2, Pack, NoOp, Placeholder, (For more information see https://docs.nvidia.com/deeplearning/dgx/tf-trt-user-guide/index.html#supported-ops).
2019-12-04 14:32:46.830196: I tensorflow/compiler/tf2tensorrt/convert/convert_graph.cc:733] Number of TensorRT candidate segments: 2
2019-12-04 14:32:47.161950: I tensorflow/stream_executor/platform/default/dso_loader.cc:42] Successfully opened dynamic library libcudart.so.10.0
2019-12-04 14:33:14.933859: I tensorflow/stream_executor/platform/default/dso_loader.cc:42] Successfully opened dynamic library libcudnn.so.7
2019-12-04 14:33:17.632085: I tensorflow/stream_executor/platform/default/dso_loader.cc:42] Successfully opened dynamic library libcublas.so.10.0

Hi,
There is a sample of converting TF to TRT:

/usr/src/tensorrt/samples/sampleUffSSD/

Please refer to it and check if it helps your usecase. Thanks.