19m 32s logged
Added image drawing and imported the drawn image to the neral network. So far the predictions are correct 40% of the time, I suspect thats because of the sharper use drawing mode compared to my training images which are rounder and more blended in yo the background. It also took me a while to realise the MNIST images have black backgrounds
19m 10s logged
Adding drawing in progress…. (will have a pygame windows where useds can draw their own digits and view the neeural networks predicion)
38m 35s logged
Touch up on the readme and installation
I build this digit classifier CLI tool, which basically allows you to pass and image and guesses which digit it features.
I built/trained the neural network from strach using pure python. I did not use and machine learning libraries because I wanted to learn how they work in a deeper level. it is currently trained on 5000 images with 15 iterations, this may sound very little but due to using python list instead of numpy arrays my runtime is alread about ~300 seconds for training.
I learned a lot about neural networks while building this project. Although I did not follow any tutorials when builing it I found 3Blue1Brown’s Naural Network videos extremely helpfull when researching fundemental concpets! Even though the research part was hard I feel like the hardest part was optimizing the train function I had to do more research on how batch gradient decesent works and more techniques.
Its currently at 92% accuracy with testing data but it might still styrugle with picture that have grey or weird back grounds which I plan to fix next!
The model can be tested using the PyPi install route, I also have detailed install/run instructions in my readme.
Thank you for reading allong that far!
57m 40s logged
I’m finally on PyPi! It took way longer than I expected to turn the trained model into teh CLI tool but here we are! I also polised and edited the README. It can be installed by python3 -m pip install digitReader and run by classifyDigit –image image.jpeg. Feel free to check out my project demo!
3h 10m 19s logged
Worked on making the woring project a comman line app that precits the digit in the given picture. Next I will write the readme
1h 4m 10s logged
Fixed my batch training, my loss calculation was off. Also added weight saving so I dont have to retarin everytime I want to test. And of course seperated test from train! Although it still runs preetry slow 50s train runtime…. I will experinment with different batch sizes, training sizes, number of neurons, learning rates soon. I might also do a version with numpy, my original goal was to not use any libraries to learn the fundementals, but numpy is much faster than python list so I might use that to reduce runtime in a seperate version to see how much of a difference it makes!
57m 31s logged
I tried to optimize and add more trainign data but accidently made it even less acurate and take longer to process. Will have to do some research!
49m 3s logged
Trained the network with 10 images (first tried with 100 but it was SO SLOW), but it runs and makes a prediction!!!!! altough it thought 5 was a 3….. I am gonna try figure out ways to optimize my train function next fo I can train 100!
44m 30s logged
Decoding MINST images! Deconding those with no libraries are harder than I thought and I had to do a TON of research about the data files. But here it is the color value of each pixel of the first imege!! I just have to extract them all, maybe find a better way to store them so I dont have to decode them each time, and pass them to my neural network, which currently have 784 hidden input nurons, 64 hidden neurons and 10 output neurons corresponding to each digit I am trying to predict. Sorry for the super long devlog, I guess I am just excited that its turning into what I am trying to build!
1h 59m 55s logged
I added 3 output neurons meaning the neural network will respond in tems of class 1: 0.01, class 2: 0.03, class 3: 0.96. To implement that I had to add softmax fucntion chnage my forwardPass and backwardPass implementations to work with new matrixes.
@BeytiKebab
on
3D-google-dino-game
26m 55s logged
tried to add the go forward logic, i think something is wrong but its getting late…
1h 54m 26s logged
Added a second varable to the hidden neurons, and organised the code to work with it. Managed to get a very low loss!!!!!
@BeytiKebab
on
beytiKebab's_Hackpad
32m 40s logged
Added silkscreen and resired the PCB
@BeytiKebab
on
beytiKebab's_Hackpad
48m 57s logged
Redid the connections on my PCB for checks to run clear, I migth add some designs next! Note: Previusly I worked on my PCB for 2 hours and 19 minutes but those hourse were accidently logged to my neural network project instead. I asked about the situation in #stardance-help, and support confirmed that the time cannot be reassigned due to Hackatime’s data protections. I mentioned the siyuation in my devlog for my neural network projects devlog too when submmiting the two hours there as adviced. To learn more: https://hackclub.slack.com/archives/C0AP0NMSP3P/p1788725484466219
2h 19m 39s logged
I worked on the PCB for my Hackpad project for 2 hours and 19 minutes, but accidentally connected the Lapce recording to this project instead. I asked about the situation in #stardance-help, and support confirmed that the time cannot be reassigned due to Hackatime’s data protections. They advised me to mention the situation in both projects’ devlogs so the extra time can be accounted for during review. More details can be found here https://hackclub.slack.com/archives/C0AP0NMSP3P/p1788725484466219
@BeytiKebab
on
beytiKebab's_Hackpad
1h 19m 5s logged
Redid the Schematic, because forgot to add the dsplay. Also included LED’s and keyboard matrix this time!
@BeytiKebab
on
beytiKebab's_Hackpad
23m 57s logged
Got started with the PCB, but realised I FORGOT the OLED screen… Will go back and do that 👍
@BeytiKebab
on
beytiKebab's_Hackpad
28m 30s logged
Did the schematic design!!
@BeytiKebab
on
3D-google-dino-game
18m 45s logged
Just added cactus obejct not need to generate them continously and add colision detection
