There are many good resources for learning computer vision. This repo is a collection of resources that I have found useful. One of the best is the OpenCV-Python Tutorials.

and This PyIageSearch blog is also very good: PyImageSearch https://github.com/BasitJaved/PyImageSearch_University

Google it you will figure it out

Check out the right command for you: [https://pytorch.org/get-started/locally/]

to create in conda env from scratch you could do this:

conda create -n pytorch torchvision -c pytorch

CONDA_SUBDIR=osx-64 conda create -n torch_36_x86 python=3.6
conda activate torch_36_x86

conda create -n torch_36 python=3.6
conda activate torch_36

This folder holds some fun I had with openCV the readme in the folder has more info

I found this simple example in OpenCV "MobileNetSSD_deploy"  https://github.com/Beomus/Python-Realtime-Object-Detector/blob/master/main.py

this holds an example of houw to us transfer learning with pytorch the inference method is not working yet

this holds an example of how to use pytorch on a camera feed

image detection example is here: https://github.com/deepsense-ai/Keras-PyTorch-AvP-transfer-learning

they use resnet50 to train on a custom dataset for alien vs predator detection

well like 5 of them... there are more added all the time it seems. Here is an example of pytorch Yolo v3.

git clone https://github.com/nrsyed/pytorch-yolov3.git

cd pytorch-yolov3
pip install .
./get_weights.sh

if you have a wget error on macos you can use brew to install wget and then run the script again

brew install wget

if you don't have brew installed you can go install it and try again