Data Science Stack Exchange is a question and answer site for Data science professionals, Machine Learning specialists, and those interested in learning more about the field. It only takes a minute to sign up.
Sign up to join this community
I'd like to get started with machine learning, specifically image recognition. I know that Python is the most popular language for ML since it's easy to pick up and there's tons of libraries for it. I have a few questions:
What are the advantages of using a pre-trained model vs training my own?
The main advantage of pre-trained models is that it greatly reduces required data for your model. You can see it as a better way to initialize your model.
Where would I find pre-trained models? At what point would it be worth it to train my own models, and how can I do so?
Keras has a variety of pretrained models for image classification. If your data consists of rare objects or uncommon situations, it might be better to train your own model, say for very specific task. Pretrained models only work if patterns in your data share similarities to patterns learned by the pretrained model.
Which library should I be using as a beginner? TensorFlow, scikit-learn, etc.
Scikit-learn and Keras are pretty straightforward. They both have toy datasets for experimenting. Deep Learning with Python by François Chollet has a nice chapter on computer vision, which has a section on using pretrained models.
Hopefully You will get your answer after open this below link.
https://www.pyimagesearch.com/2018/09/24/opencv-face-recognition/
