🐛 [Bug] Nightly import fails

[laikhtewari]

Bug Description

Installed nightly and got OSError: libnvinfer_plugin.so.10: cannot open shared object file: No such file or directory

To Reproduce

pip install -U --pre torch_tensorrt --index-url https://download.pytorch.org/whl/nightly/cu121
python3 -c "import torch_tensorrt"

Error:

Traceback (most recent call last):
  File "<string>", line 1, in <module>
  File "/home/ltewari/scratch/miniconda3/envs/nightly/lib/python3.9/site-packages/torch_tensorrt/__init__.py", line 102, in <module>
    _register_with_torch()
  File "/home/ltewari/scratch/miniconda3/envs/nightly/lib/python3.9/site-packages/torch_tensorrt/__init__.py", line 96, in _register_with_torch
    torch.ops.load_library(trtorch_dir + "/lib/libtorchtrt.so")
  File "/home/ltewari/scratch/miniconda3/envs/nightly/lib/python3.9/site-packages/torch/_ops.py", line 1242, in load_library
    ctypes.CDLL(path)
  File "/home/ltewari/scratch/miniconda3/envs/nightly/lib/python3.9/ctypes/__init__.py", line 374, in __init__
    self._handle = _dlopen(self._name, mode)
OSError: libnvinfer_plugin.so.10: cannot open shared object file: No such file or directory

Environment

Build information about Torch-TensorRT can be found by turning on debug messages

• Torch-TensorRT Version (e.g. 1.0.0): torch-tensorrt==2.4.0.dev20240429+cu121
• PyTorch Version (e.g. 1.0): torch==2.4.0.dev20240507+cu121
• CPU Architecture: x86_64
• OS (e.g., Linux): Ubuntu
• How you installed PyTorch (conda, pip, libtorch, source): pip (nightly)
• Build command you used (if compiling from source): NA
• Are you using local sources or building from archives: NA
• Python version: Python 3.9.18
• CUDA version: 12.4
• GPU models and configuration: H100 PCIe
• Any other relevant information:

[narendasan]

@laikhtewari do you have tensorrt installed?

[laikhtewari]

Looks like it, I have tensorrt==8.6.1.post1 Shouldn't that be installed automatically as a dependency when I pip install torch_tensorrt?