How to build your own neural network from scratch in Python
towardsdatascience.com
How to build your own neural network from scratch in Python
1–10 of 46 posts
Re: How to build your own neural network from scratch in Python
#2Too often you see articles like this and they start with
$ import a_whole_bunch_of_stuff
Good to see that this is not the case here :)
The fast.ai course has a similar exercise in the beginning, but you'll still import the weights from somewhere else.
Their fast.ai v1 library has a very short implementation too (loading the MINIST example dataset and then using Resnet18):
from fastai import *
from fastai.data import *
untar_data(MNIST_PATH)
data = image_data_from_folder(MNIST_PATH)
learn = ConvLearner(data, tvm.resnet18, metrics=accuracy)
learn.fit(1)
Source: http://docs.fast.ai/
Re: How to build your own neural network from scratch in Python
#3Towards Data Science is definitely one of the best Medium Publications. I hope they don't begin to monetise with Medium's member-only content. It's such a drawback to Medium.
Re: How to build your own neural network from scratch in Python
#4Thanks for submitting, this is going to come in handy for my education! I have an assignment next week which requires creating a neural network, then substituting various optimizers in place of backpropagation and comparing performance over iterations. Have found that simple numpy based NN's are easier to examine and connect the changes with the theory, this guide looks to help further with this understanding!
Re: How to build your own neural network from scratch in Python
#5Another thing to try is calculating backpropagation by hand, on paper, with a small NN. This is what my mentor said his final exam on NN in college involved.
Re: How to build your own neural network from scratch in Python
#6I can highly recommend Joel Grus’ live coding video. He creates a deep learning library only using numpy in an hour and it’s really fun to watch it all come together. https://youtu.be/o64FV-ez6Gw
Re: How to build your own neural network from scratch in Python
#7Nah, let's make that 10 lines: https://gist.github.com/vesche/72dafed33d614710f03f1b75cff1c...
I got a beer for anyone who can golf it under 5 lines
Re: How to build your own neural network from scratch in Python
#8Too often you see articles like this and they start with $ import a_whole_bunch_of_stuff Good to see that this is not the case here :) The fast.ai course has a similar exercise in the beginning, but you'll still import the weights from somewhere else. Their fast.ai v1 library has a very short implementation too (loading the MINIST example dataset and then using Resnet18): from fastai import * from fastai.data import…
How to make cake, by Russell Peters
First, you get cake. Then you make it for 20 minutes. Then you have cake.
Re: How to build your own neural network from scratch in Python
#9Nah, let's make that 10 lines: https://gist.github.com/vesche/72dafed33d614710f03f1b75cff1c... I got a beer for anyone who can golf it under 5 lines
Not readable, doesn't count ;)
Re: How to build your own neural network from scratch in Python
#10Another thing to try is calculating backpropagation by hand, on paper, with a small NN. This is what my mentor said his final exam on NN in college involved.
I would have thought calculus 1 was a prerequisite for that course.