import LightningModule error

[lxzyuan]

When I run the demo of gdl_apps/EMOCA/demos/test_emoca_on_images.py, I get the following errors:

Can u help me to solve this problem? Many thanks!

[nerlfield]

Hello! I also faced such a problem. I fixed the error by updating pytorch-lightning to a higher version (==1.5.10) and setuptools (==61.2.0)

Here is my updated requirements36.txt:

cython~=0.29
pytorch-lightning==1.5.10
adabound~=0.0.5
addict~=2.4.0
alabaster~=0.7.12
albumentations~=1.0.2
babel~=2.8.0
chumpy~=0.70
compress-pickle~=1.2.0
configparser~=5.0.1
deepspeech-gpu~=0.9.2
docker-pycreds~=0.4.0
docutils~=0.16
dominate~=2.6.0
easydict~=1.9
face-alignment~=1.3.3
facenet-pytorch~=2.5.1
flatbuffers~=1.12
freetype-py~=2.2.0
gdown~=3.12.2
gitdb~=4.0.5
gitpython~=3.1.11
google-pasta~=0.2.0
hydra-core~=1.0.5
imagesize~=1.2.0
imgaug~=0.4.0
importlib-resources~=5.1.0
insightface~=0.5
kaggle~=1.5.12
keras-preprocessing~=1.1.2
kornia~=0.6.1
lightning-bolts~=0.4.0
matplotlib~=3.3.4
memory-profiler~=0.58.0
mmcv~=1.2.2
munch~=2.5.0
numpy~=1.19.5
omegaconf~=2.0.6
onnx~=1.10.2
onnxruntime-gpu~=1.9.0
opencv-python~=4.5.1.48
opencv-python-headless~=4.5.3.56
opt-einsum~=3.3.0
pathtools~=0.1.2
pillow~=8.1.2
prettytable~=2.4.0
protobuf~=3.13.0
psutil~=5.7.3
pyasn1-modules~=0.2.8
python-speech-features~=0.6
pyzmq~=22.0.3
scikit-sound~=0.2.3
scipy~=1.5.4
sentry-sdk~=0.19.2
setuptools
shapely~=1.7.1
smmap~=3.0.4
subprocess32~=3.5.4
tb-nightly~=2.5.0a20201210
tensorboard-data-server~=0.6.1
timm~=0.4.9
torchfile>=0.1.0
torchgeometry>=0.1.2
torchmetrics>=0.5.1
trimesh~=3.6.18
wandb~=0.10.30
watchdog~=0.10.3
wrapt~=1.12.1
youtube-dl~=2021.6.6
pytorch3d @ git+https://github.com/facebookresearch/[email protected]

[lxzyuan]

Thank you very much! It works~

[radekd91]

This one is tricky. It might be machine specific, I am not sure.

One would expect that enforcing a particular lightning version should do the trick but it seems you guys had problems with it anyway. Can you please check and confirm that in fact the installed version was the one listed in the requirement file? Maybe you had a leftover lightning installation in your environment?

[lxzyuan]

Thank you @radekd91 & @animtel.

I followed @animtel's instructions and installed a higher version of pytorch-lightning(1.5.10), which has solved the problem.

By the way, can you provide the test code for the AFEW-VA dataset (maybe I didn't find it?), I would like to reproduce the results in sup. mat.

Also, when I run emoca/gdl/datasets/AfewVaDataModule.py, I get the errors: FileNotFoundError: [Errno 2] No such file or directory '.../not_processed/gt.pkl'

Can you help me to solve these problems? Best wishes to you~~