ASMR Coding - training and evaluate several CNNs | Let's learn - CNN from scratch in Go - 17
Now I trained few different CNN models and tested them against first version that was built fully on top of dense layers only.
🖥️ You can support my channel: https://ko-fi.com/roman_v
📹 Main channel: / @_romanv_
💬 Discord: / discord
✉️ Contact: [email protected]
💻 Code: https://github.com/RomanVolkov/nnfs_go
Here I continue to dive into neural networks area, now with implementation of Convolutional Neural Networks. I use the foundation from Neural networks from scratch series to build new Layers and other things for CNN. Again, all the code will be written in Golang instead of Python, but we will use Python code as references
The goal is to learn how neural networks are made inside and how it looks like to build is fully from the scratch without using frameworks like pytorch or tensorflow. We’ll have to find alternative in Go or reimplement some operations fully from the ground.
Timecodes:
0:00 - Video's goal
1:08 - Generic Model factory
13:35 - Making trainig of several models at once
19:39 - Surprised that 2-layers CNN worked as expected
22:57 - Let's start training
27:08 - Note about training speed
28:39 - Note about possible improvement to CNN
29:24 - Note CNN 2
30:43 - Results :)
31:30 - What is next
👀 Disclaimer:
Views and opinions are my own and do not represent or reflect the opinions of my current or past employer(s).
#learncoding #softwareengineering #neuralnetwork #machinelearning #ai #golang