AI417 - Assignment 4

Student Information

Organization: AI417UPM
Student ID: 4410058
Name: Retaj

Repository Contents

dummy.py → Dummy model architecture
Vanilla.py → Vanilla model architecture
dummy-weights.bin → Trained weights for Dummy model
vanilla-weights.bin → Trained weights for Vanilla model
hubconf.py → Torch Hub configuration file

Model Description

Dummy Model

The dummy model is a simple feed-forward neural network composed only of Linear (fully connected) layers.
It does not include any activation functions between layers.
Therefore, it represents a purely linear transformation.

Vanilla Model

The vanilla model is a feed-forward neural network that includes activation functions between Linear layers.
The activation functions introduce non-linearity, allowing the model to learn more complex patterns.

How to Load the Models

You can load the models using Torch Hub:

import torch

# Load Dummy Model
model = torch.hub.load("AI417UPM/A4_4410058_Retaj", "dummy", trust_repo=True)

# Load Vanilla Model
model = torch.hub.load("AI417UPM/A4_4410058_Retaj", "vanilla", trust_repo=True)