Introduction to Neural Networks in Python (what you need to know) | Tensorflow/Keras
In this video we begin by strolling by a number of the fundamentals. We take a look at why we use neural networks and the way they perform. We do an outline of community structure (enter layer, hidden layers, output layer). We speak a bit about how you select what number of hidden layers and neurons to have. We additionally take a look at hyperparameters like batch dimension, studying price, optimizers (adam), activation capabilities (relu, sigmoid, softmax), and dropout. We end the primary part of the video speaking a bit concerning the variations between keras, tensorflow, & pytorch.
Next, we bounce into some coding examples to classify information with neural nets. In this part we load in information, do some processing, construct our community, match our information to it, after which lastly consider our mannequin. The examples get extra advanced as we go alongside. Some setup directions for the coding portion of the video are discovered beneath.
I’m going to put up a comply with up video to this quickly the place we stroll by an actual world instance the place we mechanically classify pictures of arms for the sport of rock, paper, scissors. Hopefully that needs to be up about 2 weeks from now. (EDIT: half 2 has been posted, hyperlink beneath)
If you loved this video, be sure that to like & subscribe. Feel free to depart any questions in the feedback part.
Finally by Loxbeats https://soundcloud.com/loxbeats
Creative Commons — Attribution 3.0 Unported — CC BY 3.0
Free Download: http://bit.ly/FinallyLoxbeats
Music promoted by Audio Library https://youtu.be/fGquX0Te1Yo
0:00 Video overview
1:34 Why use neural networks
3:08 How neural nets work (structure fundamentals)
6:11 Hyperparameter overview (batch dimension, optimizer, dropout, studying price, epochs)
7:53 How will we select layers, neurons, & different parameters?
9:08 Why will we need an activation perform?
10:20 What activation perform ought to I exploit?
11:25 Keras vs Tensorflow vs PyTorch
12:30 Coding begins (github & setup)
14:07 Writing our first neural community (linear instance)
18:45 Selecting optimizer & loss perform (mannequin.compile)
23:45 Fitting coaching information to our mannequin (mannequin.match)
27:31 Shuffle order of coaching information
30:12 Evaluate mannequin on check information (mannequin.consider)
32:00 Example #2: Classifying quadratic information
36:06 Example #3: Classifying 6 clusters of knowledge (attempt by yourself)
41:03 Using community to predict a single information level (mannequin.predict)
43:27 Example #4: Classifying a number of labels at a time (BinaryCrossentropy loss)
55:19 Example #5: Classifying our advanced information from begin of video
59:00 Conclusion & Next steps of studying neural nets