INTERNSHIP REPORT

A report submitted in partial fulfillment of the requirements of

BACHELOR OF TECHNOLOGY
In
COMPUTER SCIENCE AND ENGINEERING
By

Name: Vivek Payla

Roll No: 1901920100331

Under
Supervision of

Sneha Bhapkar, designation: AI Lead

IoTIoT.in

(Duration: 1/07/22 – 16/11/2022)

G.L. BAJAJ INSTITUTE OF TECHNOLOGY & MANAGEMENT,

GREATER NOIDA
Affiliated to

DR. A.P.J. ABDUL KALAM TECHNICAL UNIVERSITY,

LUCKNOW

2022-2023

1
 G.L. BAJAJ INSTITUTE OF TECHNOLOGY & MANAGEMENT,
GREATER NOIDA
DEPARTMENT OF COMPUTER SCIENCE AND ENGINEERING

CERTIFICATE

This is to certify that the “Internship report” submitted by Vivek Payla

(1901920100331) is work done by him/her and submitted during 2022 – 2023
academic year, in partial fulfillment of the requirements for the award of the degree
of BACHELOR OF TECHNOLOGY in COMPUTER SCIENCE AND
ENGINEERING, at IoTIoT.in, address - IoTIoT Innovation Lab -3 IoT C.O.E,
C.O.E.P's Bhau Institute of Entrepreneurship and Leadership Beside COEP Boat
Club C.O.E.P. Shivajinagar, Pune, Maharashtra 411005

Mr. Megh Singhal Dr. Sansar Singh Chauhan

Internship Coordinator Head of the Department

2
Certificate of Internship

3
 ACKNOWLEDGEMENT

First I would like to thank Ms.Sneha Bhapkar designation: AI Lead,

company name: IotIot.in situated at IoTIoT Innovation Lab -3 IoT C.O.E,
C.O.E.P's Bhau Institute of Entrepreneurship and Leadership Beside
COEP Boat Club C.O.E.P. Shivajinagar, Pune, Maharashtra 411005 for
giving me the opportunity to do an internship within the organization.

I also would like to acknowledge all the people that worked along with me
IoTIoT.in, IoTIoT Innovation Lab -3 IoT C.O.E, C.O.E.P's Bhau Institute
of Entrepreneurship and Leadership Beside COEP Boat Club C.O.E.P.
Shivajinagar, Pune, Maharashtra 411005 with their patience and openness
they created an enjoyable working environment.

It is indeed with a great sense of pleasure and immense sense of gratitude

that I acknowledge the help of these individuals.

I pay special thanks to my Head of the Department Dr. Sansar Singh

Chauhan for his constructive criticism throughout my internship.

I would like to thank Mr. Megh Singhal, Internship coordinator for his
support and advise to get and complete internship in above said organization.

I am extremely great full to my department staff members and friends who

helped me in successful completion of this internship.

Vivek Payla
(1901920100331)

4
 ABSTRACT

In recent years’ machine learning is playing a vital role in our everyday lifelike,
it can help us to route somewhere, find something for what we aren’t aware of,
or can schedule appointments in seconds. Looking at the other side of the coin
besides machine learning Mobile phones are equivocating and competing in the
same field. If we take an optimistic view, by applying machine learning in our
mobile devices, we can make our lives better and even move society forward.
Image Classification is the most common and trending topic of machine
learning. Among several different types of models in deep learning,
MobileNetV2 Quantised Pre-trained Models have intimated high performance
on image classification which are made out of various handling layers to gain
proficiency with the portrayals of information with numerous unique levels, are
the best AI models as of late. Here, we have used a Pre-trained Quantised
MobileNetV2 and completed the experiments on the dataset called ImageNet
and image classification, and Run inference on System and finally Deployed on
Brainypi.

5
 INDEX

S.no CONTENTS Page.no

1. Introduction 7
1.1 Object 8
2. Machine Learning 9
2.1 Supervised Learning 9
2.2 Unsupervised Learning 9
3. Deep Learning 10
3.1 Neural Networking 10
4. Tensorflow 11
4.1 Tensorflow Lite 11
4.2 Tensorflow Lite Inference 12
4.2.1 Load & Run model in Python 13
5. MobileNet 14
5.1 Model Comparison 14
6. Tools & Technology Used 15
6.1 Python 15
6.2 Google Colab 15
6.3 VsCode for Python 16
6.4 Command Prompt 16
6.5 Brainypi 16
7. Project Based On My Learning 17
7.1 Implementation On Brainypi 18
8. Conclusion 19
9. References 20

6
1. INTRODUCTION

Competency of using mobile devices in today's era has made

a significant difference by its rapid increase in utilization in the
last few years which thereby does not seem to degrade over the
coming 20+ years. On the other hand, Machine learning
/Artificial Intelligence is at the summit of making the world a
new place with wider availability of data and advancements in
storage, computing and analyzing capabilities. Correctly
consolidating them, these 2 aspects can make several
opportunities and possibilities for the upcoming new
generation.
Image classification is one of the broader fields of machine
learning. Image classification is the technique to map numbers
to symbols. To classify a dataset into various distinct classes,
their relationship i.e. between data and their classes must be
understood well for this thing to get achieved the computer
must be trained well.

7
1.1 OBJECTIVES

The objective of this internship was to learn how to

Implement a TFLite image Classification Applicatiion
On Brainypi .

● To gain skills and knowledge

This internship provided us essential skills and knowledge
one requires in the field of Machine Learning & Deep Learning.
The crucial tools used during the tenure helped us in gaining
knowledge about python programming languages.

● To get field work experience

By taking this training we enhanced our knowledge in
Machine Leaning & Deep Learning and got insight in how
applications are Developed and tested on different platforms.

● To enhance our communication skills

By interacting with my Mentor and TeamMembers I got to
learn a lot. It helped me to enhance my communicative
skills and represent my work with confidence. It boosted
my confidence to develop more project based on machine
learning create some great designs just for fun.

● To link theory with practical

First we learned the theory aspect and then we put that
into practice. By doing the practical work, our concept
got clearer and it was easy to code into Python once we
got familiar with it. By putting our theoretical knowledge
into practical, coding became more fun.

● To implementing code
By learning how to code in Python it is easy to
Run inference for different images and we can make
Amendments according to ourselves and check how
The accuracy differs.

8
2. MACHINE LEARNING

Machine learning is a field of computer science that gives computers the

ability to learn without being explicitly programmed.

Arthur Samuel, an American pioneer in the field of computer gaming and

artificial intelligence, coined the term "Machine Learning" in 1959 while at
IBM. Evolved from the study of pattern recognition and computational learning
theory in artificial intelligence, machine learning explores the study and
construction of algorithms that can learn from and make predictions on data

Machine learning tasks are typically classified into two broad categories,
depending on whether there is a learning "signal" or "feedback" available to a
learning system

2.1 Supervised Learning

Supervised learning is the machine learning task of inferring a function from

labelled training data.[1] The training data consist of a set of training
examples. In supervised learning, each example is a pair consisting of an input
object (typically a vector) and a desired output value.

A supervised learning algorithm analyses the training data and produces an

inferred function, which can be used for mapping new examples. An optimal
scenario will allow for the algorithm to correctly determine the class labels for
unseen instances. This requires the learning algorithm to generalize from the
training data to unseen situations in a "reasonable" way.

2.2 Unsupervised Learning

Unsupervised learning is the machine learning task of inferring a function to

describe hidden structure from "unlabelled" data (a classification or
categorization is not included in the observations). Since the examples given to
the learner are unlabelled, there is no evaluation of the accuracy of the structure
that is output by the relevant algorithm—which is one way of distinguishing
unsupervised learning from supervised learning and reinforcement learning.

9
3. DEEP LEARNING
Deep learning is a subset of machine learning. Usually, when people use the
term deep learning, they are referring to deep artificial neural networks, and
somewhat less frequently to deep reinforcement learning.

Deep learning is a class of machine learning algorithms that:

● Use a cascade of multiple layers of nonlinear processing units for

feature extraction and transformation. Each successive layer uses
the output from the previous layer as input.
● Learn in supervised (e.g., classification) and/or unsupervised (e.g., pattern
analysis) manners.
● Learn multiple levels of representations that correspond to different
levels of abstraction; the levels form a hierarchy of concepts.
● Use some form of gradient descent for training via backpropagation.

3.1 Neural Networking

Artificial neural networks (ANNs) or connectionist systems are computing

systems inspired by the biological neural networks that constitute animal brains.
Such systems learn (progressively improve performance on) tasks by
considering examples, generally without task-specific programming

10
4. TENSORFLOW

TensorFlow is a free and open-source software library for machine learning and
artificial intelligence. It can be used across a range of tasks but has a particular
focus on training and inference of deep neural networks

Its flexible architecture allows for the easy deployment of computation across a
variety of platforms (CPUs, GPUs), and from desktops to clusters of servers to
mobile and edge devices.

pip install tensorflow command is used to install Tensorflow on the system.

4.1 Tensorflow Lite (TFLite)

TensorFlow Lite is an open-source, product ready, cross-platform deep learning

framework that converts a pre-trained model in TensorFlow to a special format
that can be optimized for speed or storage.

The special format model can be deployed on edge devices like mobiles using
Android or iOS or Linux based embedded devices like Raspberry Pi or
Microcontrollers to make the inference at the Edge.

11
4.2 TensorflowLite Inference

The term inference refers to the process of executing a TensorFlow Lite model on-
device in order to make predictions based on input data. To perform an inference with
a TensorFlow Lite model, you must run it through an interpreter. The TensorFlow
Lite interpreter is designed to be lean and fast. The interpreter uses a static graph
ordering and a custom (less-dynamic) memory allocator to ensure minimal load,
initialization, and execution latency.

TensorFlow Lite inference typically follows the following steps:

1.Loading a model
You must load the .tflite model into memory, which contains the model's
execution graph.

2. Transforming data
Raw input data for the model generally does not match the input data format
expected by the model. For example, you might need to resize an image or
change the image format to be compatible with the model.

3.Running inference
This step involves using the TensorFlow Lite API to execute the model. It
involves a few steps such as building the interpreter, and allocating tensors, as
described in the following sections.

4.Interpreting output
When you receive results from the model inference, you must interpret the
tensors in a meaningful way that's useful in your application.

12
4.2.1 Load and run a model in Python

The Python API for running an inference is provided in the tf.lite module. From
which, you mostly need only tf.lite.Interpreter to load a model and run an inference.

The following example shows how to use the Python interpreter to load a .tflite file
and run inference with random input data

Code :

13
5. MobileNet

MobileNets are CNNs, which means that they learn parameters in convolutional
kernels that are convolved across their inputs. This approach allows the network to
identify features that may indicate ‘person’-ness, ‘car’-ness and ‘neither’-ness. The
use of CNNs instead of fully connected network means the model is robust to
translations of objects in images, it maintains an explicit hierarchical representation of
features, and contains fewer parameters. The last point is the most important in this
use case as it allows high performance with fewer parameters to store on the edge
device. MobileNets futher computational efficiency by using depthwise convolutions
combined with pointwise convolutions, which allow the model to store intermediate
results, and improve their inference performance.

5.2 Model Comparison

We also present the final sizes of the unquantised and quantised models for
MobileNetV1 and MobileNetV2. Since data augmentation does not change the
number of parameters in a model and is not used at inference time, there is no
difference in size between models that use data augmentation and their
counterparts.We also observe that data augmentation barely improves the performance
of MobileNetV1. This can be explained by observing that the model variances are
low, so data augmentation is unlikely to help as generalising is not a problem. On the
other hand, our MobileNetV1 results indicate a high bias issue. We decided to
increase the size of the model (and change architecture) to MobileNetV2. We observe
that using the larger MobileNetV2 without data augmentation presents a 1.3%
improvement over the smaller MobileNetV1 without data augmentation.

14
6. Tools & Technology used
Various tools and technology were used during our internship training.
Some of them are listed below:

6.1 PYTHON
Python is a high-level, interpreted, interactive and object-oriented scripting
language. Python is designed to be highly readable. It uses English keywords
frequently where as other languages use punctuation, and has fewer syntactical
constructions than other languages.
Python is interpreted: Python is processed at runtime by the interpreter.
You do not need to compile your program before executing it. This is
similar to Perl and PHP.

Python is Interactive: You can actually sit at a Python prompt and interact
with the interpreter directly to write your programs.

Python is Object-Oriented: Python supports Object-Oriented style

or technique of programming that encapsulates code within objects.

Python is a Beginner's Language: Python is a great language for the

beginner-level programmers and supports the development of a wide range
of applications from simple text processing to WWW browsers to games.

6.2 GOOGLE COLAB

Colab is a free Jupyter notebook environment that runs entirely in the cloud.
Most importantly, it does not require a setup and the notebooks that you create
can be simultaneously edited by your team members - just the way you edit
documents in Google Docs. Colab supports many popular machine learning
libraries which can be easily loaded in your notebook.
As a programmer, you can perform the following using Google Colab.
• Write and execute code in Python
• Document your code that supports mathematical equations
• Create/Upload/Share notebooks
• Import/Save notebooks from/to Google Drive
• Import/Publish notebooks from GitHub
• Import external datasets e.g. from Kaggle
• Integrate PyTorch, TensorFlow, Keras, OpenCV
• Free Cloud service with free GPU

15
6.3 VS CODE FOR PYTHON

working with Python in Visual Studio Code, using the Microsoft Python
extension, is simple, fun, and productive. The extension makes VS Code an
excellent Python editor, and works on any operating system with a variety of
Python interpreters. It leverages all of VS Code's power to provide auto
complete and IntelliSense, linting, debugging, and unit testing, along with the
ability to easily switch between Python environments, including virtual and
conda environments.

6.4 COMMAND PROMPT

A command prompt is the input field in a text-based user interface screen for an
operating system (OS) or program. The prompt is designed to elicit an action.
The command prompt consists of a brief text string followed by a blinking
cursor, which is where the user types command prompt commands.

Command-line interfaces (CLI) and prompts were the standard interface for
computers from the early days of computing into the 1980s. Microsoft MS-DOS
systems and other early consumer-based computers used CLIs. Current
Windows systems offer the CLI for administrative tasks. The CLI is still an
essential part of the Linux OS.

6.5 BRAINYPI

An Enterprise grade device for AI on Edge and IoT needs.It is a betterment over
Raspberrypi, more information is not provided because of privacy.

16
7. Project Based On My Learning

The Project was carried out in the following steps:-

• Found out all available TFLite Image Classification applications

available on internet.
• Choosing an application that gives accurate results.
• Running inference and comparing results
• Downloading the inferred jupyter or colab notebook.
• Running the downloaded file on local system.
• Establishing connection to the Brainypi Server.
• Finally implementing the python file on brainypi.

VS CODE Output.

VS CODE Output.

17
7.1 Implementation on Brainypi Server

Establishing Connection to Server

Implementing Python File on Brainypi

18
8. CONCLUSION

In conclusion, I can say that this internship was a greatexperience. Thanks to

this project, I acquired deeperknowledge concerning my technical skills, but I
also personally benefited.We achieve an accuracy greater than 90% on with a
model size of 690KB and latency of 0.013s. This allows us to build a haptic
garment capable of performing fast inference on the edge. With more time, we
would want to clean the dataset by hand to improve labelling. We would also
explore lower level representations of the model, such as running inference in
C++ rather than in Python, which should further improve latency.I learned to
live in a different environment from the one I amused to. Indeed, I grew more
independent in work and also in everyday life. I realized that I could do more
things than Ithought, like learning new things by myself.There are huge
opportunities available for the students whowant to work in this field.

19
9. REFRENCES

• https://www.tensorflow.org/lite/guide/inference
• https://www.tensorflow.org/lite/models/trained
• https://helloworld.co.in/article/image-classification-tensorflow-lite
• https://levelup.gitconnected.com/custom-image-classification-model-
using-tensorflow-lite-model-maker-68ee4514cd45
• https://androidapps-development-blogs.medium.com/image-
classification-android-app-with-tensorflow-lite-for-beginner-
a793655f5a0a
• https://www.tensorflow.org/lite/models/modify/model_maker

20