Computer Vision is FUN!
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
Install miniconda
YOU NEED THIS IN YOUR LIFE 😱
Google it you will figure it out
Intall pyTorch
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
Notes on Conda
For M1 macs if you need 3.6
CONDA_SUBDIR=osx-64 conda create -n torch_36_x86 python=3.6
conda activate torch_36_x86
For Intel macs and the rest of the world
conda create -n torch_36 python=3.6
conda activate torch_36
OpenCV Fun
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
PyTorch
tf_learning folder
this holds an example of houw to us transfer learning with pytorch the inference method is not working yet
torch video folder
this holds an example of how to use pytorch on a camera feed
PyTorch Detection Transfer learning example
One of the best examples 👍
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
One Yolo to rule them all
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