search

m-tassano / fastdvdnet

FastDVDnet: A Very Fast Deep Video Denoising algorithm
MIT License
572 stars 123 forks source link

Training environment on ubuntu20.04+CUDA110 #56

Closed Ssakura-go closed 1 year ago

Ssakura-go commented 1 year ago

Due to forget the environment configure before, I spared a lots time on it. Therefore, make such a log here.

The problem we may meet with:

  1. No module named cv2 Reason: opencv is not match with the cuda. Solution: config the envs by such command (pip install torch==1.7.1+cu110 torchvision==0.8.2+cu110 torchaudio==0.7.2 -f https://download.pytorch.org/whl/torch_stable.html, which is followed by the cuda version), and replace the config manner in pip with conda (such as conda install opencv)
  2. Modulenotfounderror: no module named 'skimage.measure.simple_metrics' Solution: replace the line 22 in util.py with from skimage.metrics import peak_signal_noise_ratio as compare_psnr
m-tassano commented 1 year ago

Hi, thanks for the information. You're welcome to open a pull request--there's no need to install torchaudio btw.

Ssakura-go commented 1 year ago

Hi, thanks for the information. You're welcome to open a pull request--there's no need to install torchaudio btw.

Roger that. Thank you for your reply!