Professional Documents
Culture Documents
Bachelor of Engineering
in
Computer Science and Engineering
Submitted by
Saransh Rajpurohit: (Roll No. 21UCSE4028)
Bachelor of Engineering
in
Computer Science and Engineering
Submitted by
Saransh Rajpurohit: (Roll No. 21UCSE4028)
I, Saransh Rajpurohit hereby declare that this work titled “Machine Learning” is a
record of original work done by me under the supervision and guidance of Dr. Bharat
Gupta, Coordinator and CI, E&ICT Academy at NIT Patna from 6th June to 15th July
2022.
I, further certify that this work has not formed the basis for the award of the
Degree/Diploma/Associateship/Fellowship or similar recognition to any candidate of any
university and no part of this report is reproduced as it is from any other source without
appropriate reference and permission.
SIGNATURE OF STUDENT
(Saransh Rajpurohit)
VIII Semester, CSE
Enroll. - 20R/43805
Roll No. - 21UCSE4028
3
ACKNOWLEDGEMENT
In this internship I had a great chance for learning and professional development. I
I would like to thank Dr. Bharat Gupta, my internship guide who took time out to hear,
guide, and keep me on the correct path. I am also greatly thankful to Dr. Shrwan Ram,
the Head of the Department of Computer Science and Engineering, and other faculty
4
ABSTRACT
This report encapsulates an immersive internship journey at NIT Patna, emphasizing Data
Science development. Collaborating with senior developers, I honed skills and gained
industry insights. The project, a handwritten digit recognizer application created using
Deep Learning, demonstrated technical prowess. Key features included user action
identification, task management, and team collaboration. Through agile methodologies,
continuous integration, and feedback cycles, I navigated real-world challenges.
Reflections encompass acquired skills, teamwork, and future aspirations. This experience
laid a strong foundation for professional growth, fostering a commitment to ongoing
learning and advancement in the dynamic field of software engineering.
5
Contents
INTRODUCTION 1
TECHNOLOGY USED 3
2.1 Python 3
2.2 TensorFlow 4
2.3 Slack 4
2.4 VS Code 5
2.5 Summary 5
PROJECT/WORK DETAILS 6
3.6 Conclusion 11
RESULTS/OUTCOME 12
References 14
6
List of Figures
1.1 EICT Logo .……………………………………………………………. 1
2.1 Python Logo .………………………………………………………………... 3
2.2 Slack Logo .………………………………………………………………... 4
2.3 VS Code Logo .…………………………………………………………….. 5
3.1 Input………………………………………………………………………… 6
7
Practical Training Report (B.E.)
Chapter 1
INTRODUCTION
Chapter 1: Introduction 1
Practical Training Report (B.E.)
Under the guidance of experienced data scientists and mentors, my role encompassed the
following key responsibilities:
Chapter 1: Introduction 2
Practical Training Report (B.E.)
Chapter 2
TECHNOLOGY USED
This chapter into the powerful combination of Neural Networks, Python, Tensorflow, and
Slack that fuels data applications at EICT NIT Patna. Python for core functionalities and
TensorFlow for building the deep learning model. Python's versatility facilitates data
manipulation and script automation, while TensorFlow offers the building blocks and
training capabilities to create an accurate models. Slack fosters seamless communication
and collaboration throughout the development process, ensuring all stakeholders are
aligned.
2.1 Python
2.2 TensorFlow
Tensorflow, an open-source library developed by Google, sits at the core of this project's
deep learning capabilities. TensorFlow excels at numerical computations and building
large-scale machine learning models. In this Handwritten Digit Recognition system,
TensorFlow serves two key purposes:
2. Model Training: TensorFlow facilitates the training process, where the model
learns to identify patterns and relationships within the handwritten digit image
dataset. This training process involves feeding the dataset images and their
corresponding action labels into the model, allowing it to refine its ability to
accurately predict actions from new unseen images.
By leveraging TensorFlow's powerful capabilities, the system can effectively learn from
data and achieve a high degree of accuracy in recognizing human actions from images.
2.3 Slack
2.4 VS Code
Visual Studio Code (VS Code) is a lightweight and extensible code editor developed by
Microsoft, favoured by developers for its versatility, performance, and rich ecosystem of
extensions. At EICT NIT Patna, VS Code is the preferred code editor for frontend
development, offering features such as syntax highlighting, code completion, and
integrated Git support. With a user-friendly interface and powerful customization options,
VS Code enhances developer productivity and facilitates a seamless development
experience for ML projects.
2.5 Summary
Chapter 3
PROJECT/WORK DETAILS
In this chapter, we delve into the specifics of the project undertaken during the internship
at EICT NIT Patna. In this project I focused on building a Handwritten Digit Recognizer
(HDR) system. This system analyses handwritten digit through picture or camera data to
identify and categorize specific numbers. The project involved tasks like data acquisition,
model training, and potentially building a user interface for interaction. Throughout this
process, I collaborated with experts to gain valuable insights into the complexities of
Image Recognition technology and its real-world applications. This experience not only
honed my technical skills in this specific area but also broadened my understanding of
the field and its potential impact.
The Handwritten Digit Recognizer aims to develop a system that can analyse images
containing handwritten digits and accurately identify the digit depicted. The system will
take an image as input and output the corresponding digit label (0-9). To achieve this, a
deep learning model will be trained on a large dataset of labelled handwritten digit
images, enabling the system to learn the characteristic patterns of each digit and recognize
them effectively.
• Decipher unclear digits: The system goes beyond perfectly formed digits,
aiming to recognize even unclear or messy handwriting.
• Discover hidden digits: Uncover numbers embedded within images that might
be difficult to spot with the naked eye.
• Data Access: Google Drive is mounted within the project for dataset access using
Pandas.
• Data Preprocessing: Data is divided into separate DataFrames for features and
labels. Normalization with Keras reduces training time.
• Model Training: Keras optimizers, metrics, and TensorFlow power the training
process on the designated dataset.
• Model Evaluation: The trained model's performance is assessed using the testing
dataset.
• User Input: OpenCV's cv2.imread reads the user's image and converts it to
grayscale (28x28 pixels).
• Prediction: NumPy converts the image to an array, which is fed into the neural
network for digit prediction.
• While Colab isn't a traditional deployment platform, it offers several options for
sharing and running your trained model:
By separating and strategically deploying the different parts of the system, we ensure a
robust, accessible, and user-friendly Handwritten Digit Recognition experience.
Collaborating with senior data scientists and team members from diverse backgrounds
fostered the development of collaboration, communication, and teamwork skills.
Engaging in team activities such as stand-up meetings, code reviews, and sprint planning
sessions enhanced the ability to work effectively within a team environment and
contribute to shared project goals.
Exposure to industry practices and challenges provided valuable insights into the broader
landscape of software development. Learning how to navigate project requirements,
manage timelines, and prioritize tasks within a professional setting equipped with
practical knowledge and skills essential for future career growth and success.
3.6 Conclusion
The development and deployment of the human action recognizer web application on the
Neural Networks platform represented a significant learning experience and milestone in
the internship journey at EICT NIT Patna. Through hands-on experience, collaboration
with experienced developers, and exposure to industry practices, invaluable skills were
acquired, setting the foundation for continued growth and success in software
development endeavours.
Chapter 4
RESULTS/OUTCOME
The culmination of the internship experience at EICT NIT Patna has yielded tangible
results and valuable outcomes that contribute to both personal and professional growth.
Exposure to industry practices and challenges has provided valuable insights into the
broader landscape of software development. Understanding project requirements,
managing timelines, and prioritizing tasks within a professional setting have equipped me
with practical knowledge and skills essential for navigating the complexities of the
software development industry.
Chapter 4: Results/Outcomes 12
Practical Training Report (B.E.)
Chapter 5
As the journey in software development continues, the experiences gained during the
internship at EICT NIT Patna will serve as a solid foundation upon which to build and
navigate the ever-evolving landscape of technology with confidence, curiosity, and a
commitment to continuous learning and growth.
References
[1] https://www.tensorflow.org/
[2] https://www.ibm.com/topics/computer-vision
[3] https://www.sas.com/en_in/insights/analytics/computer-vision.html
[4] https://aws.amazon.com/what-is/neural-
network/#:~:text=A%20neural%20network%20is%20a,inspired%20by%20the%20
human%20brain.
[5] https://news.mit.edu/2017/explained-neural-networks-deep-learning-0414
[6] https://pillow.readthedocs.io/
[7] https://realpython.com/image-processing-with-the-python-pillow-library/
[8] https://slack.com/
[9] https://code.visualstudio.com/
References 14