Neural Network from Scratch - Machine Learning Python

From scratch implementation of neural network in Python using only the numpy library. There's a lot of mathematics behind Neural Networks particularly for back-prop and I've made previous videos going through the mathematics for NN and a blog post to enable us to focus on the implementation in this video.

❤️ Support the channel ❤️

Paid Courses I recommend for learning (affiliate links, no extra cost for you):

✨ Free Resources that are great:

💻 My Deep Learning Setup and Recording Setup:

GitHub Repository:

✅ One-Time Donations:

▶️ You Can Connect with me on:

[AladdinPersson]

Just a small correction in the video I said 3 layered where I was counting the input layer as well: this is not correct, as you don't normally count the input layer. This implementation is then of a 2 layered Neural net.

[abderemanesoilihi]

Thank you really much. The explanation is really clear. very informative

[vygintasmarciulaitis]

What type of input does this NN expects? 3 numbers corresponding to x, y of an dot and third coresponding to color?

[usmanshakeel]

does the reg_loss representing l1 or l2 regularization method? because I am not sure about this. And what does 0.5 represents in the formula of reg_loss

[youssefhany]

please please please explain the point net archeticture and what does each layer do ( input transform, feature transform, mlp, max pool) please.

[pratikhmanas]

I just have a question. Can we make the number of layers to be added parameterised?

[bryanchambers]

Hello, I'm wondering why you recommended that Michael Nielson page. I wasted like three days on it, after going through the algorithms he tells you to import things that don't exist like:
>>import network. Same thing with the way he imports nmist library. Its nothing but error messages and frustration. His SGD algorithm is garbage too, nothing but error messages.

[alonalon]

where is the math you said you go through in the first minute? you are talking there about previous videos. I don't find them.
thanks in advance and thanks for all your videos.

[juank]

Could you help: How can I assign the accuracy to a variable, like a=accuracy or a = binary_accuracy. This is part of the code of the neural network:


model = Sequential()



model.add(Dense(48, input_dim=48, activation='relu'))



model.add(Dense(24, activation='relu'))



model.add(Dense(2, activation='sigmoid'))

model.compile(loss='mean_squared_error',



optimizer='adam',



metrics=['binary_accuracy'])



model.fit(training_data, target_data, epochs=1000)

scores = model.evaluate(training_data, target_data)



"training_data, target_data are arrays"

Result of evaluation:

binary_accuracy: 0.5000



binary_accuracy: 50.00%

[codeforcoders]

Make a video on how to classify a image using a logistics regression in python without using TF, keras