TypeError: pybind11::init(): factory function returned nullptr.

[chegnyanjun]

Hi,i am use tensorrt7.0.0.11 and cuda10.1 ubuntu16.04 pytorch1.3.1 python3.6.But i faced above problem when i was using it.It indices the problem from this line:

TRT_LOGGER = trt.Logger(trt.Logger.WARNING)
with trt.Builder(TRT_LOGGER) as builder, builder.create_network() as 
 network, trt.OnnxParser(network,TRT_LOGGER) as parser:  #<--- this line got above problem.

The code can be use in the tensorrt6.0.1.5,but got problem in the tensorrt7.0.0.11.

What's problem of my code?

[rmccorm4]

Hi @chegnyanjun,

I don't think there's anything wrong with that code snippet, I use it all the time. However, I notice you're using CUDA 10.1, and I believe TRT7 only supports 9.0, 10.0, and 10.2 per the support matrix: https://docs.nvidia.com/deeplearning/sdk/tensorrt-support-matrix/index.html

Please try again after changing to a supported CUDA version or use a container if you don't want to change your host install

[chegnyanjun]

＠rmccorm4,thanks for your help,i will try it.

[chegnyanjun]

＠rmccorm4,Hi,i have tried your suggestion, and i change my cuda10.1 to cuda10.0, but it still has such problem,the whole problem is:
[TensorRT] ERROR: CUDA initialization failure with error 35. Please check your CUDA installation: http://docs.nvidia.com/cuda/cuda-installation-guide-linux/index.html
Traceback (most recent call last):
master/scripts/Pytorch2ONNX.py", line 53, in ONNX2Engine
with trt.Builder(TRT_LOGGER) as builder:
TypeError: pybind11::init(): factory function returned nullptr
Can you help me? tkx.

[taomiao]

met the same problem. @rmccorm4
Titan xp
nvidia-driver 418.56
cuda 10.0

[ligonzheng]

met the same problem.
cuda 10.2
tensorrt 7.0.0.11

[dmenig]

Same problem and same version as @ligonzheng .

[TensorRT] ERROR: CUDA initialization failure with error 1. Please check your CUDA installation:  http://docs.nvidia.com/cuda/cuda-installation-guide-linux/index.html
[2020-02-04 11:54:00,016] PID : 5850 : CRITICAL : 
Traceback (most recent call last):
  File "file.py", line 278, in get_engine
    with open(engine_file_path, "rb") as f, trt.Runtime(TRT_LOGGER) as runtime:
 Uncaught exception: pybind11::init(): factory function returned nullptr

[rmccorm4]

@chegnyanjun @taomiao @ligonzheng @hyperfraise

You all got this error:

[TensorRT] ERROR: CUDA initialization failure with error 35. Please check your CUDA installation: http://docs.nvidia.com/cuda/cuda-installation-guide-linux/index.html

1. Did you verify that your CUDA install is good as the error suggests?

   • Can you run nvidia-smi successfully?
   • Can you run any sample GPU apps in PyTorch/TensorFlow?

2. Does your TensorRT installation actually match your current CUDA installation? For example @chegnyanjun you mention that you downgraded from CUDA 10.1 to CUDA 10.0, but did you also switch to the TensorRT CUDA 10.0 release as well? Or did you install the TensorRT CUDA 10.2 release? There are several permutations to choose from here: https://developer.nvidia.com/nvidia-tensorrt-7x-download

Example:

• CUDA 10.0 .tar.gz package for Ubuntu: https://developer.nvidia.com/compute/machine-learning/tensorrt/secure/7.0/7.0.0.11/tars/TensorRT-7.0.0.11.Ubuntu-18.04.x86_64-gnu.cuda-10.0.cudnn7.6.tar.gz

vs.

• CUDA 10.2 .tar.gz package for Ubuntu: https://developer.nvidia.com/compute/machine-learning/tensorrt/secure/7.0/7.0.0.11/tars/TensorRT-7.0.0.11.Ubuntu-18.04.x86_64-gnu.cuda-10.2.cudnn7.6.tar.gz

The TensorRT package you install has to match your environment (OS, CUDA version, CUDNN version, etc.)

[rawar]

I've got the same error with Google Colab.

!dpkg -l | grep TensorRT

ii graphsurgeon-tf 7.0.0-1+cuda10.2 amd64 GraphSurgeon for TensorRT package ii libnvinfer-bin 7.0.0-1+cuda10.2 amd64 TensorRT binaries ii libnvinfer-dev 7.0.0-1+cuda10.2 amd64 TensorRT development libraries and headers ii libnvinfer-doc 7.0.0-1+cuda10.2 all TensorRT documentation ii libnvinfer-plugin-dev 7.0.0-1+cuda10.2 amd64 TensorRT plugin libraries ii libnvinfer-plugin7 7.0.0-1+cuda10.2 amd64 TensorRT plugin libraries ii libnvinfer-samples 7.0.0-1+cuda10.2 all TensorRT samples ii libnvinfer7 7.0.0-1+cuda10.2 amd64 TensorRT runtime libraries ii libnvonnxparsers-dev 7.0.0-1+cuda10.2 amd64 TensorRT ONNX libraries ii libnvonnxparsers7 7.0.0-1+cuda10.2 amd64 TensorRT ONNX libraries ii libnvparsers-dev 7.0.0-1+cuda10.2 amd64 TensorRT parsers libraries ii libnvparsers7 7.0.0-1+cuda10.2 amd64 TensorRT parsers libraries ii python3-libnvinfer 7.0.0-1+cuda10.2 amd64 Python 3 bindings for TensorRT ii python3-libnvinfer-dev 7.0.0-1+cuda10.2 amd64 Python 3 development package for TensorRT ii tensorrt 7.0.0.11-1+cuda10.2 amd64 Meta package of TensorRT ii uff-converter-tf 7.0.0-1+cuda10.2 amd64 UFF converter for TensorRT package

!nvidia-smi

`Sun May 3 16:56:38 2020
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 440.64.00 Driver Version: 418.67 CUDA Version: 10.1 |
|-------------------------------+----------------------+----------------------+
| GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. |
|===============================+======================+======================|
| 0 Tesla P100-PCIE... Off | 00000000:00:04.0 Off | 0 |
| N/A 37C P0 25W / 250W | 0MiB / 16280MiB | 0% Default |
+-------------------------------+----------------------+----------------------+

+-----------------------------------------------------------------------------+
| Processes: GPU Memory |
| GPU PID Type Process name Usage |
|=============================================================================|
| No running processes found |
+-----------------------------------------------------------------------------+`

When I run

with trt.Builder(TRT_LOGGER) as builder, builder.create_network(*EXPLICIT_BATCH) as network, trt.OnnxParser(network, TRT_LOGGER) as parser: builder.max_workspace_size = 1 << 28 builder.max_batch_size = 1 builder.fp16_mode = True

I've got

`TypeError Traceback (most recent call last)
in ()
----> 1 with trt.Builder(TRT_LOGGER) as builder, builder.create_network(*EXPLICIT_BATCH) as network, trt.OnnxParser(network, TRT_LOGGER) as parser:
2 builder.max_workspace_size = 1 << 28
3 builder.max_batch_size = 1
4 builder.fp16_mode = True

TypeError: pybind11::init(): factory function returned nullptr`