search

shunsukesaito / SCANimate

This repository contains the code for the paper "SCANimate: Weakly Supervised Learning of Skinned Clothed Avatar Networks"
https://scanimate.is.tue.mpg.de/
Other
274 stars 31 forks source link

'smpl' has no attribut 'create' #26

Closed ireneisme closed 1 year ago

ireneisme commented 1 year ago

Thanks for sharing the code as well as the instruction to implemente it.

I tried to set up the environment on the colab platform but failed to run the test code with the error showed below. I wonder how to fix this problem, I'd appreciate any possible solution!!!

Traceback (most recent call last):
  File "/usr/local/envs/scanimate/lib/python3.8/runpy.py", line 194, in _run_module_as_main
    return _run_code(code, main_globals, None,
  File "/usr/local/envs/scanimate/lib/python3.8/runpy.py", line 87, in _run_code
    exec(code, run_globals)
  File "/content/SCANimate/apps/test_scanimate.py", line 167, in <module>
    testWrapper()
  File "/content/SCANimate/apps/test_scanimate.py", line 164, in testWrapper
    test(opt, args.test_dir)
  File "/content/SCANimate/apps/test_scanimate.py", line 53, in test
    model = smpl.create(opt['data']['smpl_dir'], model_type='smpl_vitruvian',
AttributeError: module 'smpl' has no attribute 'create'
ireneisme commented 1 year ago

Seems like I didn't activate the virtual environment before set up the smpl model, so the version of python and cuda... is incorrect. Now the test run successfully!! By the way, I use a higer version for the whole environment(python3.8 pytorch-cuda=11.6 pytorch=1.13.0) because I cannot install those package listed in the intall.sh on colab!

By the way, I also set up the environment on ubuntu, and encounter a similar problem when setup smpl model. The reason is my 'python3' is not using the correct python version which I installed for the virtual environment. More detail can be found https://stackoverflow.com/questions/53372272/python-venv-virual-environment-uses-wrong-version-of-python