SqueezeNet-Implementation

This repository attempts to replicate the SqueezeNet architecture using TensorFlow discussed in the research paper: "SqueezeNet: AlexNet-level accuracy with 50x fewer parameters and <0.5MB model size".

The paper can be read here.

The official implementation of this paper can be found here.

Requirements

1. tensorflow-gpu==1.13.1 ( Does not work with Tensorflow 2.x)
2. sklearn
3. opencv-python
4. numpy
5. Python 3.x ( Specifically not python 3.8, anything else works)

Architecture Implemented

1. Fire Module

2. SqueezeNet Module

Working

The data used for this implementation was picked up from the Kaggle Dataset - Soil Types

• Step 1: Clone the repository

git clone https://github.com/RohanMathur17/SqueezeNet-Implementation.git

• Step 2: Install necessary libraries as discussed in Requirements section
• Step 3: Within train.py, change your path for data at line 31

Change this line 
base_dir = '/content/gdrive/MyDrive/SqueezeNet/data/'

• Step 4: In your command prompt, run the train.py file to train the model

python train.py

Additional Information

• This repository attempts to replicate the architecture only. Performance may vary based on parameters implemented. Can change the same and experiment using the train.py module.
• A sample usage of this can be found in the Notebook here.