This is a collection of my personal projects that I work on in my free time. Hope you like them.
Flash
An end-to-end Deep Learning platform that allows users to create, train, and deploy their own neural network models in a matter of minutes without writing a single line of code.
The platform supports two types of tasks:
-
Image Classification
Classify images from your own dataset by using them to train a ResNet-34 or MobileNet v2 model. Training happens via transfer learning where models available will be pre-trained on the ImageNet dataset. -
Sentiment Analysis
Predict sentiment from sentences by training a LSTM or GRU based sequential model on your own dataset. The models will be trained from scratch.
Tools Used: Python, PyTorch, JavaScript, ReactJS, Redux, AWS (Lambda, EC2, S3)
TensorNet
A high-level deep learning library built on top of PyTorch to solve computer vision problems.
Tools Used: Python, PyTorch
Topic Based Image Captioning
An automatic image caption generation system built using Deep Learning.
- Devised a novel image captioning model using CNNs and LSTMs trained on MSCOCO-2017 Dataset.
- Created a system where LSTMs were given additional information (topics) extracted from image captions using Latent Dirichlet Allocation (LDA).
Tools Used: Python, Tensoflow, NLTK, OpenCV
Stock Bridge
An online stock market simulator enabling users to have an experience of trading in the real-world market.
Tools Used: Python, Django, Django REST Framework, Bootstrap, Heroku, sendgrid
Code Warrior
An online judge platform with support for languages C, C++, and Python.
Tools Used: Python, Django, Bootstrap, Amazon Web Services, PythonAnywhere, sendgrid.
Kart
An E-commerce website built using Django.
- Built the backend on entirely on Django. Utilized jQuery to make the website asynchronous.
- Developed features like checkout with online payment, send order receipt via email, selling digital items e.t.c.
Tools Used: Python, Django, Bootstrap, jQuery, Ajax, jsrender, stripe, mailchimp, Amazon Web Services, heroku, sendgrid
Autoranking Amazon Reviews
Ranking the reviews on Amazon according to their helpfulness score.
- The problem was modeled as a regression problem. The performance was evaluated by using the coefficient of determination and rank correlation.
- Predictions were made based on various categories of features of the review text, and other metadata associated with the review, with the purpose of generating a rank for a given list of reviews.
Tools Used: Python, Numpy, Pandas, textblob, scikit-learn