Installation issue

[drozzy]

I am doing something wrong. When I follow the proposed install instructions:

conda create --name codd python=3.8 -y
conda activate codd
pip install scipy pyyaml terminaltables natsort
pip install torch==1.12.1+cu113 torchvision==0.13.1+cu113 torchaudio==0.12.1 --extra-index-url https://download.pytorch.org/whl/cu113 # pytorch
pip install --no-index --no-cache-dir pytorch3d -f https://dl.fbaipublicfiles.com/pytorch3d/packaging/wheels/py38_cu113_pyt1121/download.html # pytorch3d

The last command causes:

ERROR: Could not find a version that satisfies the requirement fvcore (from pytorch3d) (from versions: none)
ERROR: No matching distribution found for fvcore

I'm on ubuntu 18.04.

[drozzy]

Fixing this with pip install fvcore.

[drozzy]

pip install git+https://github.com/princeton-vl/lietorch.git fails with:

 The detected CUDA version (10.2) mismatches the version that was used to compile
      PyTorch (11.3). Please make sure to use the same CUDA versions.

      [end of output]

  note: This error originates from a subprocess, and is likely not a problem with pip.
  ERROR: Failed building wheel for lietorch
  Running setup.py clean for lietorch
Failed to build lietorch
ERROR: Could not build wheels for lietorch, which is required to install pyproject.toml-based projects

[mli0603]

Your CUDA version is older than the instructions.

Instead of

pip install torch==1.12.1+cu113 torchvision==0.13.1+cu113 torchaudio==0.12.1 --extra-index-url https://download.pytorch.org/whl/cu113 # pytorch

Can you try

pip install torch==1.12.1+cu102 torchvision==0.13.1+cu102 torchaudio==0.12.1 --extra-index-url https://download.pytorch.org/whl/cu102

[drozzy]

That seems to have fixed the lietorch. Now when I try:

python benchmark_speed.py configs/models/codd.py

/home/adrozdyuk/.conda/envs/codd/lib/python3.8/site-packages/mmcv/__init__.py:20: UserWarning: On January 1, 2023, MMCV will release v2.0.0, in which it will remove components related to the training process and add a data transformation module. In addition, it will rename the package names mmcv to mmcv-lite and mmcv-full to mmcv. See https://github.com/open-mmlab/mmcv/blob/master/docs/en/compatibility.md for more details.
...
    import mmengine
ModuleNotFoundError: No module named 'mmengine'

[mli0603]

Hi @drozzy

You did not install mmcv, probably mmseg too

pip install mmcv-full==1.7.0 -f https://download.openmmlab.com/mmcv/dist/cu102/torch1.12/index.html
pip install mmsegmentation 

[drozzy]

I ran both of those. Same result. I'll try again from the beginning.

[drozzy]

Unfortunately I cannot get this to work. This is the error that occurs every time. I'll close this for now.

(codd) adrozdyuk-dt:CODD$ python benchmark_speed.py configs/models/codd.py
/home/adrozdyuk/.conda/envs/codd/lib/python3.8/site-packages/mmcv/__init__.py:20: UserWarning: On January 1, 2023, MMCV will release v2.0.0, in which it will remove components related to the training process and add a data transformation module. In addition, it will rename the package names mmcv to mmcv-lite and mmcv-full to mmcv. See https://github.com/open-mmlab/mmcv/blob/master/docs/en/compatibility.md for more details.
  warnings.warn(
Traceback (most recent call last):
  File "benchmark_speed.py", line 9, in <module>
    from model.builder import build_estimator
  File "/home/adrozdyuk/.../codd/CODD/model/__init__.py", line 3, in <module>
    from .builder import *
  File "/home/adrozdyuk/.../codd/CODD/model/builder.py", line 5, in <module>
    from mmseg.models.builder import MODELS
  File "/home/adrozdyuk/.conda/envs/codd/lib/python3.8/site-packages/mmseg/__init__.py", line 5, in <module>
    import mmengine
ModuleNotFoundError: No module named 'mmengine'

[drozzy]

I can't even install pip requirements as indicated. For example, install of fvcore is required to make it work for me (before I get further errors).

Can someone try creating the conda environment from scratch and following the pip install instructions in the readme? Because in my experience, conda should be relatively bulletproof.

[mli0603]

I wonder if this is due to Ubuntu 18.04. I have not tested systems with CUDA 10.2 or earlier.