Debanjan Chaudhuri (Deep)'s starred repositories
Bat-Ball-Tracking-System
Real-time Object Detector with YOLOv5
bangla-tts
Bangla text to speech, Multilingual (Bangla, English) real-time ([almost] in a GPU) speech synthesis library
bengaliAnalyzer
This module helps to analyze Bengali sentences. It can analyze various entities. Can do non contextual PoS tagging. Is capable of returning the lemmas present in a sentence.
BanglaSpeech2Text
BanglaSpeech2Text: An open-source offline speech-to-text package for Bangla language. Fine-tuned on the latest whisper speech to text model for optimal performance.
Bangla-Punctuation-Corrector
A project for punctuation restoration from Bangla text.
Contextual-Spell-Checker-For-Bangla
Automatic Context Sensitive Spelling Correction for Bangla Text Using Bert and Levenstein Distance
Bangla-stemmer
A Python package to get stem of any inflected Bangla words.
Data-Augmentation-Engineering-Drawing
Repo for engineering drawing synthesis
supervision
We write your reusable computer vision tools. 💜
neural-networks-and-deep-learning
Code samples for my book "Neural Networks and Deep Learning"
RelationPrompt
This repository implements our ACL Findings 2022 research paper RelationPrompt: Leveraging Prompts to Generate Synthetic Data for Zero-Shot Relation Triplet Extraction. The goal of Zero-Shot Relation Triplet Extraction (ZeroRTE) is to extract relation triplets of the format (head entity, tail entity, relation), despite not having annotated data for the test relation labels.
pdf-redactor
A general purpose PDF text-layer redaction tool for Python 2/3.
zeroshot_topics
Topic Inference with Zeroshot models
spaCy-entity-linker
spaCy module for linking text to Wikidata items
Optimizing-Windmill-layout-Competition-by-Shell-Ltd
Shell.ai Hackathon for Sustainable and Affordable Energy The world needs to move to a cleaner energy system if it is to meet growing energy demand while tackling climate change. In April 2020, Shell shared its ambition to become a net-zero emissions energy business by 2050, or sooner. Renewable electricity is central to this ambition. Electricity is the fastest-growing part of the energy system and, when generated from renewable sources such as wind, has a big role to play in reducing greenhouse gas emissions. We see digitalisation and AI as key enablers to the energy transition. Challenge: Windfarm Layout Optimisation In this Shell.ai Hackathon for Sustainable and Affordable Energy, we invite you to optimise the placement of 50 wind turbines of 100 m height and100 m rotor diameter each on a hypothetical 2D offshore wind farm area such that the AEP (Annual Energy Production) of the farm is maximized. One of the key problems of an unoptimized layout is the combined effect wind turbines can have on the wind speed distribution in a windfarm. As a wind turbine extracts energy from incoming wind, it creates a region behind it downstream where the wind speed is decreased- this is called a wake region. Note that wind turbines automatically orient their rotors, to face incoming wind from any direction. Due to the induced speed deficit, a turbine placed inside the wake region of an upstream turbine will naturally generate reduced electrical power. This inter-turbine interference is known as a wake effect. An optimal windfarm layout is important to ensure a minimum loss of power during this combined wake effect. This Shell.ai Hackathon for Sustainable and Affordable Energy edition, focuses on an interesting and complex coding problem. When competing, you will face challenges such as a high dimensionality, complex multimodality and the discontinuous nature of the search space. This makes optimizing the layout analytics difficult. But, armed with optimization strategies and computer algorithms, you can solve this problem.
DeepFaceLive
Real-time face swap for PC streaming or video calls