- 1. Introduction
- 2. Resources
- 3. Advice
- 4. Ending Note
Hey y'all. Vincent here. I'll keep this document short (I'll need water just from typing this). I give resources and advice here and a small ending note.
Deep Learning
- Deep Learning Crash Course for Beginners by freeCodeCamp covers just the conceptual basics of deep neural networks.
- Keras with TensorFlow Course - Python Deep Learning and Neural Networks for Beginners Tutorial by deeplizard provided by freeCodeCamp covers deep learning with a concentration on computer vision models. It focuses on practical implementation (with TensorFlow and Keras) rather than math.
- Introduction | Deep Learning Tutorial 1 (Tensorflow Tutorial, Keras & Python) by codebasics covers a myriad of topics, covering concepts and their implementations in TensorFlow and Keras.
- TensorFlow 2.0 Complete Course - Python Neural Networks for Beginners Tutorial by TechWithTim provided by freeCodeCamp briefly covers concepts while focusing mainly on the code implementation in TensorFlow and Keras.
- Hands-On Machine Learning with Scikit-Learn, Keras, and TensorFlow by Aurélien Géron is a comprehensive book for breaking into machine learning and deep learning. Part II of the book covers deep learning with TensorFlow and Keras.
- PyTorch Tutorials by Aladdin Persson offers a code-centric guide to learning deep learning with the PyTorch framework.
For those who are a bit rusty (or are new to deep learning), definitely check out deeplizard's channel (which covers both code and concepts). Specifically, check out deeplizard's series Deep Learning Fundamentals - Intro to Neural Networks for the conceptual overview of deep learning and check out TensorFlow - Python Deep Learning Neural Network API to learn the deep learning library TensorFlow.
Faster R-CNN
There is a family of special CNN architectures suited for the task of object detection called R-CNNs or Region-based CNN.
- R-CNN
- Fast R-CNN
- Faster R-CNN
As you can probably tell from the names of these R-CNNs, each one is faster in terms of speed than the previous one. I will provide mostly resources for
Faster R-CNN, but I will also have additional resources on R-CNN and Fast R-CNN.
R-CNN
Fast R-CNN
Faster R-CNN
- paperswithcode is a website for machine learning papers and code. In the link I provided, it gives the paper and some popular repos that implement the faster r-cnn.
- chenyuntc's implementation on a Simplified (Clean Code) Version of the Faster R-CNN.
- Paperspace blog on Faster R-CNN is pretty thorough (covers also R-CNN and Fast R-CNN).
- analytics-vidhya blog is really thorough (covers r-CNN and Fast R-CNN).
This list is not comprehensive. When in doubt, google or search on YouTube! But these were the resources I found that I thought were pretty thorough.
Mask R-CNN
None of these are comprehensive. Cross compare them, experiment with Mask R-CNN prebuilt code (which you won't use in the actual project), and lastly Google more if you ever find yourself lost.
- Interact with your learning. Experiment with and question everything.
- Don't be discouraged! The field can be daunting if you're a beginner. If you have any questions or concerns reach out to your mentors or Google!
- Meet with the team often! This is how we make a collaborative effort in progressing in the project. We can also hash out project details, logistics, and other things.
Good luck and have fun on the project guys. I hope y'all learn a lot. If you have any questions (which you will) ask us (me and Jon) or try Googling! Reach out to me for further questions about how to get caught up with the material for the project. I can closely guide you through the shortest path to getting up to speed in deep learning. Well that's it for this document, I hope it was helpful. Looking forward to making an awesome project with you guys (and even more excited for you guys to learn more about this field).