this is a deep learning workshop done in Wameedh Scientific CLub.
We Started by:
1.learning the Concept of Machine Learning throughout linear regression and logistic regression. 2.Introducing the different Machine Learning algorithms like : decision tree, random forest , Support vector machine and some codes. 3.we introduced numpy, pandas, scikit-learn libraries followed by linear regression full explanation with code and introducing neural networks; and giving a task in logistic regression. 4.Starting with deep learning by introducing deep learning algorithms like: deep neural networks, Recurrent neural networks, Long short term memory, Transformer model, Generative adversial networks , and Convolutional neural networks. 5.diving deep into Convolutional neural networks by explaining each layer and introducing Pytorch framework. 6.simple project code demonstration of VGG19 classification model for flower classification; and a task to be done classification.diving deep into the concepts of Machine learning and deep learning, i recommend andrew ng courses on deeplearning.ai : Machine Learning Specialization : https://www.deeplearning.ai/courses/machine-learning-specialization/ Deep Learning specialization : https://www.deeplearning.ai/courses/deep-learning-specialization/
To have a good visual understanding of neural networks, i recommend : https://www.youtube.com/watch?v=aircAruvnKk&list=PLZHQObOWTQDNU6R1_67000Dx_ZCJB-3pi
for the ones who want accelerated learning, you can see the MIT course with labs : http://introtodeeplearning.com/ with corresponding videos : https://www.youtube.com/playlist?list=PLTZ1bhP8GBuTCqeY19TxhHyrwFiot42_U
Scikit-learn map : https://scikit-learn.org/stable/tutorial/machine_learning_map/index.html
virtual environements are very important in handling dependencies, for each AI project you need some certain dependencies, so , if you have many projects, there will be conflicts and errors in dependencies , for example if project1 requires pytorch V1.0 and project2 requires pytorch V2.0 ==> THIS WILL RESULT AN ERROR or conflict. in that case we use virtual environement so that we can install for each project its requirements and versions.
conda
venv you can use either one of them. PS: venv is faster than conda.pandas DataFrame : https://www.w3schools.com/python/pandas/pandas_dataframes.asp
pytorch layers and function : https://pytorch.org/docs/stable/nn.html
VGG19 paper : https://arxiv.org/pdf/1409.1556.pdf