pytorch_quantization/cuda_ext.cpython-38-x86_64-linux-gnu.so: undefined symbol: _ZN3c1015SmallVectorBaseIjE8grow_podEPvmm

[Charmve]

the error log:

root@d1a0e04b40c1:/Lidar_AI_Solution/CUDA-BEVFusion# python qat/export-camera.py --ckpt=model/resnet50int8/bevfusion_ptq.pth
Traceback (most recent call last):
  File "qat/export-camera.py", line 40, in <module>
    from pytorch_quantization.nn.modules.tensor_quantizer import TensorQuantizer
  File "/opt/conda/lib/python3.8/site-packages/pytorch_quantization/__init__.py", line 20, in <module>
    from .quant_modules import *
  File "/opt/conda/lib/python3.8/site-packages/pytorch_quantization/quant_modules.py", line 23, in <module>
    from pytorch_quantization import nn as quant_nn
  File "/opt/conda/lib/python3.8/site-packages/pytorch_quantization/nn/__init__.py", line 19, in <module>
    from pytorch_quantization.nn.modules.tensor_quantizer import *
  File "/opt/conda/lib/python3.8/site-packages/pytorch_quantization/nn/modules/tensor_quantizer.py", line 24, in <module>
    from pytorch_quantization.tensor_quant import QuantDescriptor, tensor_quant, fake_tensor_quant, scaled_e4m3
  File "/opt/conda/lib/python3.8/site-packages/pytorch_quantization/tensor_quant.py", line 28, in <module>
    from pytorch_quantization import cuda_ext
ImportError: /opt/conda/lib/python3.8/site-packages/pytorch_quantization/cuda_ext.cpython-38-x86_64-linux-gnu.so: undefined symbol: _ZN3c1015SmallVectorBaseIjE8grow_podEPvmm

some try:

root@d1a0e04b40c1:/Lidar_AI_Solution/CUDA-BEVFusion# pip uninstall pytorch_quantization
root@d1a0e04b40c1:/Lidar_AI_Solution/CUDA-BEVFusion# pip install pytorch_quantization
Collecting pytorch_quantization
  Using cached pytorch-quantization-2.2.0.tar.gz (6.8 kB)
  Preparing metadata (setup.py) ... error
  error: subprocess-exited-with-error
  
  × python setup.py egg_info did not run successfully.
  │ exit code: 1
  ╰─> [16 lines of output]
      Traceback (most recent call last):
        File "<string>", line 2, in <module>
        File "<pip-setuptools-caller>", line 34, in <module>
        File "/tmp/pip-install-yxsyqxxd/pytorch-quantization_b7d5804e946d41759b83610b4467a0de/setup.py", line 137, in <module>
          raise RuntimeError(open("ERROR.txt", "r").read())
      RuntimeError:
      ###########################################################################################
      The package you are trying to install is only a placeholder project on PyPI.org repository.
      This package is hosted on NVIDIA Python Package Index.
      
      This package can be installed as:
      ```
      $ pip install --no-cache-dir --extra-index-url https://pypi.nvidia.com pytorch-quantization
      ```
      ###########################################################################################
      
      [end of output]
  
  note: This error originates from a subprocess, and is likely not a problem with pip.
error: metadata-generation-failed

× Encountered error while generating package metadata.
╰─> See above for output.

note: This is an issue with the package mentioned above, not pip.
hint: See above for details.
root@d1a0e04b40c1:/Lidar_AI_Solution/CUDA-BEVFusion# pip install --no-cache-dir --extra-index-url https://pypi.nvidia.com pytorch-quantization
Looking in indexes: https://pypi.org/simple, https://pypi.nvidia.com
Collecting pytorch-quantization
  Downloading https://pypi.nvidia.com/pytorch-quantization/pytorch_quantization-2.2.0-cp38-cp38-linux_x86_64.whl (2.9 MB)
     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 2.9/2.9 MB 774.4 kB/s eta 0:00:00
Requirement already satisfied: numpy in /opt/conda/lib/python3.8/site-packages (from pytorch-quantization) (1.23.0)
Requirement already satisfied: absl-py>=0.7.0 in /opt/conda/lib/python3.8/site-packages (from pytorch-quantization) (2.0.0)
Requirement already satisfied: scipy in /opt/conda/lib/python3.8/site-packages (from pytorch-quantization) (1.10.1)
Requirement already satisfied: sphinx-glpi-theme in /opt/conda/lib/python3.8/site-packages (from pytorch-quantization) (0.3)
Requirement already satisfied: prettytable in /opt/conda/lib/python3.8/site-packages (from pytorch-quantization) (3.9.0)
Requirement already satisfied: pyyaml in /opt/conda/lib/python3.8/site-packages (from pytorch-quantization) (6.0.1)
Requirement already satisfied: wcwidth in /opt/conda/lib/python3.8/site-packages (from prettytable->pytorch-quantization) (0.2.8)
Installing collected packages: pytorch-quantization
Successfully installed pytorch-quantization-2.2.0
WARNING: Running pip as the 'root' user can result in broken permissions and conflicting behaviour with the system package manager. It is recommended to use a virtual environment instead: https://pip.pypa.io/warnings/venv
root@d1a0e04b40c1:/Lidar_AI_Solution/CUDA-BEVFusion# 

the error has been living:

root@d1a0e04b40c1:/Lidar_AI_Solution/CUDA-BEVFusion# python -c "import pytorch_quantization as pq; print(pq.__version__)"
Traceback (most recent call last):
  File "<string>", line 1, in <module>
  File "/opt/conda/lib/python3.8/site-packages/pytorch_quantization/__init__.py", line 20, in <module>
    from .quant_modules import *
  File "/opt/conda/lib/python3.8/site-packages/pytorch_quantization/quant_modules.py", line 23, in <module>
    from pytorch_quantization import nn as quant_nn
  File "/opt/conda/lib/python3.8/site-packages/pytorch_quantization/nn/__init__.py", line 19, in <module>
    from pytorch_quantization.nn.modules.tensor_quantizer import *
  File "/opt/conda/lib/python3.8/site-packages/pytorch_quantization/nn/modules/tensor_quantizer.py", line 24, in <module>
    from pytorch_quantization.tensor_quant import QuantDescriptor, tensor_quant, fake_tensor_quant, scaled_e4m3
  File "/opt/conda/lib/python3.8/site-packages/pytorch_quantization/tensor_quant.py", line 28, in <module>
    from pytorch_quantization import cuda_ext
ImportError: /opt/conda/lib/python3.8/site-packages/pytorch_quantization/cuda_ext.cpython-38-x86_64-linux-gnu.so: undefined symbol: _ZN3c1015SmallVectorBaseIjE8grow_podEPvmm

[hopef]

Hi, wei,

You can refer the official site of pytorch_quantization_library to check it.

------------------ Original ------------------ From: Wei ZHANG @.> Date: Thu,Oct 26,2023 6:49 PM To: NVIDIA-AI-IOT/Lidar_AI_Solution @.> Cc: Subscribed @.***> Subject: Re: [NVIDIA-AI-IOT/Lidar_AI_Solution] pytorch_quantization/cuda_ext.cpython-38-x86_64-linux-gnu.so: undefined symbol: _ZN3c1015SmallVectorBaseIjE8grow_podEPvmm (Issue #207)

the error log: @.:/Lidar_AI_Solution/CUDA-BEVFusion# python qat/export-camera.py --ckpt=model/resnet50int8/bevfusion_ptq.pth Traceback (most recent call last): File "qat/export-camera.py", line 40, in <module> from pytorch_quantization.nn.modules.tensor_quantizer import TensorQuantizer File "/opt/conda/lib/python3.8/site-packages/pytorch_quantization/init.py", line 20, in <module> from .quant_modules import * File "/opt/conda/lib/python3.8/site-packages/pytorch_quantization/quant_modules.py", line 23, in <module> from pytorch_quantization import nn as quant_nn File "/opt/conda/lib/python3.8/site-packages/pytorch_quantization/nn/init.py", line 19, in <module> from pytorch_quantization.nn.modules.tensor_quantizer import * File "/opt/conda/lib/python3.8/site-packages/pytorch_quantization/nn/modules/tensor_quantizer.py", line 24, in <module> from pytorch_quantization.tensor_quant import QuantDescriptor, tensor_quant, fake_tensor_quant, scaled_e4m3 File "/opt/conda/lib/python3.8/site-packages/pytorch_quantization/tensor_quant.py", line 28, in <module> from pytorch_quantization import cuda_ext ImportError: /opt/conda/lib/python3.8/site-packages/pytorch_quantization/cuda_ext.cpython-38-x86_64-linux-gnu.so: undefined symbol: _ZN3c1015SmallVectorBaseIjE8grow_podEPvmm
some try: @.:/Lidar_AI_Solution/CUDA-BEVFusion# pip uninstall pytorch_quantization @.:/Lidar_AI_Solution/CUDA-BEVFusion# pip install pytorch_quantization Collecting pytorch_quantization Using cached pytorch-quantization-2.2.0.tar.gz (6.8 kB) Preparing metadata (setup.py) ... error error: subprocess-exited-with-error × python setup.py egg_info did not run successfully. │ exit code: 1 ╰─> [16 lines of output] Traceback (most recent call last): File "<string>", line 2, in <module> File "<pip-setuptools-caller>", line 34, in <module> File "/tmp/pip-install-yxsyqxxd/pytorch-quantization_b7d5804e946d41759b83610b4467a0de/setup.py", line 137, in <module> raise RuntimeError(open("ERROR.txt", "r").read()) RuntimeError: ########################################################################################### The package you are trying to install is only a placeholder project on PyPI.org repository. This package is hosted on NVIDIA Python Package Index. This package can be installed as: $ pip install --no-cache-dir --extra-index-url https://pypi.nvidia.com pytorch-quantization ########################################################################################### [end of output] note: This error originates from a subprocess, and is likely not a problem with pip. error: metadata-generation-failed × Encountered error while generating package metadata. ╰─> See above for output. note: This is an issue with the package mentioned above, not pip. hint: See above for details. @.:/Lidar_AI_Solution/CUDA-BEVFusion# pip install --no-cache-dir --extra-index-url https://pypi.nvidia.com pytorch-quantization Looking in indexes: https://pypi.org/simple, https://pypi.nvidia.com Collecting pytorch-quantization Downloading https://pypi.nvidia.com/pytorch-quantization/pytorch_quantization-2.2.0-cp38-cp38-linux_x86_64.whl (2.9 MB) ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 2.9/2.9 MB 774.4 kB/s eta 0:00:00 Requirement already satisfied: numpy in /opt/conda/lib/python3.8/site-packages (from pytorch-quantization) (1.23.0) Requirement already satisfied: absl-py>=0.7.0 in /opt/conda/lib/python3.8/site-packages (from pytorch-quantization) (2.0.0) Requirement already satisfied: scipy in /opt/conda/lib/python3.8/site-packages (from pytorch-quantization) (1.10.1) Requirement already satisfied: sphinx-glpi-theme in /opt/conda/lib/python3.8/site-packages (from pytorch-quantization) (0.3) Requirement already satisfied: prettytable in /opt/conda/lib/python3.8/site-packages (from pytorch-quantization) (3.9.0) Requirement already satisfied: pyyaml in /opt/conda/lib/python3.8/site-packages (from pytorch-quantization) (6.0.1) Requirement already satisfied: wcwidth in /opt/conda/lib/python3.8/site-packages (from prettytable->pytorch-quantization) (0.2.8) Installing collected packages: pytorch-quantization Successfully installed pytorch-quantization-2.2.0 WARNING: Running pip as the 'root' user can result in broken permissions and conflicting behaviour with the system package manager. It is recommended to use a virtual environment instead: https://pip.pypa.io/warnings/venv @.:/Lidar_AI_Solution/CUDA-BEVFusion#
the error has been living: @.:/Lidar_AI_Solution/CUDA-BEVFusion# python -c "import pytorch_quantization as pq; print(pq.version)" Traceback (most recent call last): File "<string>", line 1, in <module> File "/opt/conda/lib/python3.8/site-packages/pytorch_quantization/init.py", line 20, in <module> from .quant_modules import * File "/opt/conda/lib/python3.8/site-packages/pytorch_quantization/quant_modules.py", line 23, in <module> from pytorch_quantization import nn as quant_nn File "/opt/conda/lib/python3.8/site-packages/pytorch_quantization/nn/init.py", line 19, in <module> from pytorch_quantization.nn.modules.tensor_quantizer import * File "/opt/conda/lib/python3.8/site-packages/pytorch_quantization/nn/modules/tensor_quantizer.py", line 24, in <module> from pytorch_quantization.tensor_quant import QuantDescriptor, tensor_quant, fake_tensor_quant, scaled_e4m3 File "/opt/conda/lib/python3.8/site-packages/pytorch_quantization/tensor_quant.py", line 28, in <module> from pytorch_quantization import cuda_ext ImportError: /opt/conda/lib/python3.8/site-packages/pytorch_quantization/cuda_ext.cpython-38-x86_64-linux-gnu.so: undefined symbol: _ZN3c1015SmallVectorBaseIjE8grow_podEPvmm
— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you are subscribed to this thread.Message ID: @.***>

[adidier17]

I wonder if this is a version issue. @hopef can you share what is the version number of pytorch_quantization that you used?

@Charmve you have downloaded version 2.20, which in the release history was released September 28 2023, but the latest commit in this repo is from July 16. However if you pull the available versions using pip directly

pip index --extra-index-url https://pypi.nvidia.com versions pytorch-quantization
WARNING: pip index is currently an experimental command. It may be removed/changed in a future release without prior warning.
pytorch-quantization (2.2.1)
Available versions: 2.2.1, 2.2.0, 2.1.3, 2.1.2

version 2.1.2, which is not listed on the site page is available. I do not know its date of release, so perhaps it is correct? @hopef can you confirm if this is the version that was used or is there possibly another older version that is no longer available?

[W6WM9M]

I encountered the same error with pytorch-quantization=2.2.1 but the error is resolved with pytorch-quantization=2.1.3

[bigsheep2018]

update: 2.1.3 pre-build version not work. build from source works.