Skip to content
/ MATML Public

SGD, MAML and MATML algorithms along with toy datasets

Notifications You must be signed in to change notification settings

h-0-0/MATML

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

5 Commits
 
 
 
 

Repository files navigation

MATML

Code used for my masters project on "Can machine learn?" Inlcudes code for stochastic gradient descdent (SGD), Model-Agnostic-Meta-Learning (MAML) and Model-Agnostic-Time-adjusted-Meta-Learning (MATML) algorithms. These algorithms are explained in detail in the report for my project also included in the repo. It also can carry out k-shot learning able to calculate alpha as explained in the report for both MAML and MATML.

Alongside these algorithms it also includes a class to generate neural networks and data from toy datasets including sin curve regression (one for generating sin curves with different amplitude and phase one for generating sin curves with a phase related to the time at which the sin curve is found) and binary gaussian classification (where the distributions shift over time).

It also includes many plotting functions alongside the ability to choose activation function, loss function, optimiser, data structure, etc...

About

SGD, MAML and MATML algorithms along with toy datasets

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages