varungupta31 / dashcam_anonymizer

Code to Blur Human Faces and Vehicle License Plates in Video and Images using a SoTA Object Detection model YOLOv8
MIT License
23 stars 2 forks source link

Generation of environment.yml for mac #2

Closed binho closed 1 year ago

binho commented 1 year ago

I'm trying the project on mac and I'm getting error:

(base) ➜  dashcam_anonymizer git:(master) ✗ conda env create -f environment.yml
Collecting package metadata (repodata.json): done
Solving environment: failed

ResolvePackageNotFound:
  - x264=1!157.20191217
  - libvpx=1.7.0
  - libclang=10.0.1
  - libgfortran5=13.1.0
  - libstdcxx-ng=11.2.0
  - blas=1.1
  - ld_impl_linux-64=2.38
  - _libgcc_mutex=0.1
  - libgcc-ng=11.2.0
  - libgomp=11.2.0
  - libgfortran-ng=13.1.0
  - icu=58.2
  - _openmp_mutex=5.1
  - libxkbcommon=1.0.1

Is it possible to export a mac version with suggested command here? https://stackoverflow.com/a/58160918

Thank you

varungupta31 commented 1 year ago

@binho Hey, thanks for your interest in this project. I see, I'm yet to find a concrete way to create completely reliable platform independent conda environment configs.

You may just git pull or git clone the repo again.

I have updated the repo now with a new Readme, and some other fixes, hope it solved the issue for you. Just create from the new .yaml file, and try again.

Note: Be careful processing long videos. In my M1, I got a segmentation fault parsing a 1500 frame video, so I guess a GPU would be required at a certain point, if not finding a way to optimise memory in the script.

If you still face issue in the environment (perhaps due to OS variations):

  1. Create a conda environment with python.
  2. pip install the opencv package [pip install opencv-python==4.6.0.66 recommended]
  3. Pip install pybboxes package.
  4. Pip install the Ultralytics package [pip install ultralytics] and you're all set!

Close the issue, if it works for you now.

binho commented 1 year ago

Thank you! Env file worked. I see the issue you mentioned regarding seg fault. I'll try to find something over here too 👍