Me + Rhees learing how to make a neural network

Good article on how to do it from scratch "https://www.kaggle.com/code/soham1024/basic-neural-network-from-scratch-in-python"

---

Where we learned it

This video and this github is what I followed:

• https://github.com/lukepolson/youtube_channel/blob/main/Python%20Tutorial%20Series/pytorch4.ipynb
• https://www.youtube.com/watch?v=gBw0u_5u0qU&t=607s

---

Questions

1. Where are these random numbers coming from and do they matter? (batch_size = 5, epoch = 20, lr=0.01) etc
2. Understading the MyNeuralNet class more (why 3 matrix's?, understanding the foward pass more?)
3. dont know what squeeze or view does
4. understanding why everything is a tensor
5. RELU
6. The cross entropy loss function
7. Stochastic gradient descecnt
8. why zero out the gradient

---

Anserws

1. lr=0.01 stands for learning rate and is a hyper paremeter in Stochastic gradient descecnt