alexpasqua / nns-from-scratch Goto Github PK
View Code? Open in Web Editor NEWImplementation from scratch of fully connected neural networks trainable through SGD with backpropagation
License: GNU General Public License v3.0
Implementation from scratch of fully connected neural networks trainable through SGD with backpropagation
License: GNU General Public License v3.0
Bug loss function when batch size is less than 5
Documentation on some of those scripts scripts is missing / bad
In the main
section of the functions scripts (acts & losses), the objects in the dictionary aren't used, we're calling directly the functions but for testing purposes, at this point, it doesn't make sense to do so
Ho provato ad aggiungere la regularization tramite la funzione che creammo.
La funzione che abbiamo creato, concettualmente è la seguente:
lambda * sum ( weights) **2
Mi è sorto un dubbio. Questa formula è valida quando viene aggiunta alla loss prima di essere derivata, quindi prima di passare dal ciclo di back.
Nel momento in cui dobbiamo usarla però, a noi ci serve 2lambdaw
la somma dovrebbe essere a livello teorico, ma quando noi aggiorniamo i pesi w, non dobbiamo considerare la somma del vettore pesi, ma soltanto il valore w che vogliamo aggiornare. Giusto?
Ad esempio
update w1
w1 = w1 + eta * grad_w1 -lambda * w1 <--- giusto? Non dobbiamo mica considerare la somma di w1,w2 qui
When weight_init is set to "random" the Network is unable to see the values passed as upper_lim and lower_lim.
It always return random values in range (0, 1).
As backprop is set right now, there is no weight update in the last layer.
MUST be fixed
Aggiungere Target da terminale.
Currently the script does not work if executed with arguments, they're not parsed properly.
bug in network.py
setter weights in Layer doesn't work... needs to be checked
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google ❤️ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.