CodeFormer
CodeFormer copied to clipboard
would the script be Apple M1 friendly?
PyTorch since supported M1 GPU with device = MPS , would you consider put the option into script?
I tried to use it on my M1 MBP image processing seem work, but during video processing output path doesn't have any result.
Hi @WebberShaw , we have fixed this issue, you can pull the new update and check out our new commands for video enhancement. Note that:
# For Windows/Mac users, please install ffmpeg first
conda install -c conda-forge ffmpeg
If it still doesn't work, please re-install ffmpeg-python by running:
pip uninstall ffmpeg-python
pip install ffmpeg-python
@chicman That sounds great, I will take a look at it and try to include it in the next update.
@chicman Have updated to support M1 chip now!
Great. That all apple silicone gpu can work on this now. On Jan 21, 2023 at 14:52 +0800, Shangchen Zhou @.***>, wrote:
@chicman Have updated to support M1 chip now! — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you were mentioned.Message ID: @.***>
Just tried the latest master on M1 gpu, here I got some missing modules.
(torch-gpu) @.*** CodeFormerv3 % python inference_codeformer.py -w 0.5 --bg_upsampler realesrgan --input_path ./inputs/cropped_faces
Traceback (most recent call last):
File "/Users/jm/miniconda3/CodeFormerv3/inference_codeformer.py", line 7, in
@chicman Have updated to support M1 chip now! — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you were mentioned.Message ID: @.***>
OK never mind, just executed the basic/setup.py fixed the problem!
M1 GPU now works! On Jan 23, 2023 at 6:56 AM +0800, Juehming Shi @.***>, wrote:
Just tried the latest master on M1 gpu, here I got some missing modules.
(torch-gpu) @.*** CodeFormerv3 % python inference_codeformer.py -w 0.5 --bg_upsampler realesrgan --input_path ./inputs/cropped_faces Traceback (most recent call last): File "/Users/jm/miniconda3/CodeFormerv3/inference_codeformer.py", line 7, in
from basicsr.utils import imwrite, img2tensor, tensor2img File "/Users/jm/miniconda3/CodeFormerv3/basicsr/init.py", line 11, in from .version import gitsha, version ModuleNotFoundError: No module named 'basicsr.version' On Jan 21, 2023 at 2:52 PM +0800, Shangchen Zhou @.***>, wrote: @chicman Have updated to support M1 chip now! — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you were mentioned.Message ID: @.***>
Dears, I want to ask how to use the m1 gpu in this project.
I tried these, and able to accelerate with Apple Silicon (M1):
cd CoderFormer python3 ./basicr/setup.py install