AES (1) AI (1) Attention (1) Autoencoders (1) BERT (1) CNN (2) Course Project (3) DNN (2) Data Science - Visual Search - CNN - Image Classification - Image Segmentation (1) Deep Image Prior (1) Deep Language Model (2) Deep Learning (6) Deep Learning Accelerators (1) Digital Communication (1) Distributed Processing (2) Embedded Systems (1) Energy Efficient Accelerators (1) FPGA (1) Industry Collaboration (2) IoT - Smart Home - ESP8266 - Embedded - AWS - EC2 - Mood Light (1) LLMs (2) Large Language Models (1) MIMO (1) ML Acceleration (4) ML Accelerators (1) Machine Learning Acceleration (1) Machine Learning Acclerator Design (1) Major Project (1) Melanoma - Skin Cancer - MultiResUnet - Skin Segmentation - Image Classification (1) NLP (5) Parallel Computing (1) Personalized Recommendation (2) RNN (1) SGD (1) SIH (1) Safety (1) Smart India (1) Software-Hardware Codesign (2) Transformers (2) VLSI Design (1) Wireless Communications (1) course assignment (3) data mining (3) hackathon (1) hardware (1) machine learning (1) numpy (1)

 AES (1)

LiqVid <> SNU: Two Stage Automated Essay Scoring (AES)

 AI (1)

The Art of Defending

 Attention (1)

SpAtten: Efficient Sparse Attention Architecture with Cascade Token and Head Pruning

 Autoencoders (1)

Massive MIMO Channel Estimation using Deep Neural Networks

 BERT (1)

LiqVid <> SNU: Two Stage Automated Essay Scoring (AES)

 CNN (2)

Distributed Deep Learning Using Synchronous Stochastic Gradient Descent
Eyeriss: An Energy-Efficient Reconfigurable Accelerator for Deep Convolutional Neural Networks

 Course Project (3)

Self Balancing Bike
Analog VLSI implementation of Support Vector Machine Learning and Classification
Working on Li-Fi

 DNN (2)

RecNMP: Accelerating Personalized Recommendation with Near-Memory Processing
Centaur: A Chiplet-based, Hybrid Sparse-Dense Accelerator for Personalized Recommendations

 Data Science - Visual Search - CNN - Image Classification - Image Segmentation (1)

ViSenze Project Report

 Deep Image Prior (1)

Massive MIMO Channel Estimation using Deep Neural Networks

 Deep Language Model (2)

Contributing to the Bhashini Initiative
LiqVid <> SNU: Two Stage Automated Essay Scoring (AES)

 Deep Learning (6)

HyPar: Towards Hybrid Parallelism for Deep Learning Accelerator Array
Distributed Deep Learning Using Synchronous Stochastic Gradient Descent
SpAtten: Efficient Sparse Attention Architecture with Cascade Token and Head Pruning
Switch Transformers: Scaling to Trillion Parameter Models with Simple and Efficient Sparsity
Eyeriss: An Energy-Efficient Reconfigurable Accelerator for Deep Convolutional Neural Networks
Contributing to the Bhashini Initiative

 Deep Learning Accelerators (1)

Eyeriss: An Energy-Efficient Reconfigurable Accelerator for Deep Convolutional Neural Networks

 Digital Communication (1)

Working on Li-Fi

 Distributed Processing (2)

HyPar: Towards Hybrid Parallelism for Deep Learning Accelerator Array
Distributed Deep Learning Using Synchronous Stochastic Gradient Descent

 Embedded Systems (1)

Self Balancing Bike

 Energy Efficient Accelerators (1)

Eyeriss: An Energy-Efficient Reconfigurable Accelerator for Deep Convolutional Neural Networks

 FPGA (1)

Centaur: A Chiplet-based, Hybrid Sparse-Dense Accelerator for Personalized Recommendations

 Industry Collaboration (2)

Contributing to the Bhashini Initiative
LiqVid <> SNU: Two Stage Automated Essay Scoring (AES)

 IoT - Smart Home - ESP8266 - Embedded - AWS - EC2 - Mood Light (1)

Bisquare Project Report

 LLMs (2)

Switch Transformers: Scaling to Trillion Parameter Models with Simple and Efficient Sparsity
The Art of Defending

 Large Language Models (1)

Switch Transformers: Scaling to Trillion Parameter Models with Simple and Efficient Sparsity

 MIMO (1)

Massive MIMO Channel Estimation using Deep Neural Networks

 ML Acceleration (4)

HyPar: Towards Hybrid Parallelism for Deep Learning Accelerator Array
Distributed Deep Learning Using Synchronous Stochastic Gradient Descent
SpAtten: Efficient Sparse Attention Architecture with Cascade Token and Head Pruning
Switch Transformers: Scaling to Trillion Parameter Models with Simple and Efficient Sparsity

 ML Accelerators (1)

Eyeriss: An Energy-Efficient Reconfigurable Accelerator for Deep Convolutional Neural Networks

 Machine Learning Acceleration (1)

RecNMP: Accelerating Personalized Recommendation with Near-Memory Processing

 Machine Learning Acclerator Design (1)

Centaur: A Chiplet-based, Hybrid Sparse-Dense Accelerator for Personalized Recommendations

 Major Project (1)

Massive MIMO Channel Estimation using Deep Neural Networks

 Melanoma - Skin Cancer - MultiResUnet - Skin Segmentation - Image Classification (1)

SkiNet - Skin Cancer Detection and Classification

 NLP (5)

SpAtten: Efficient Sparse Attention Architecture with Cascade Token and Head Pruning
Switch Transformers: Scaling to Trillion Parameter Models with Simple and Efficient Sparsity
The Art of Defending
Contributing to the Bhashini Initiative
LiqVid <> SNU: Two Stage Automated Essay Scoring (AES)

 Parallel Computing (1)

HyPar: Towards Hybrid Parallelism for Deep Learning Accelerator Array

 Personalized Recommendation (2)

RecNMP: Accelerating Personalized Recommendation with Near-Memory Processing
Centaur: A Chiplet-based, Hybrid Sparse-Dense Accelerator for Personalized Recommendations

 RNN (1)

Word Predictor Using RNN

 SGD (1)

Distributed Deep Learning Using Synchronous Stochastic Gradient Descent

 SIH (1)

Smart India Hackathon - 2019

 Safety (1)

The Art of Defending

 Smart India (1)

Smart India Hackathon - 2019

 Software-Hardware Codesign (2)

RecNMP: Accelerating Personalized Recommendation with Near-Memory Processing
Centaur: A Chiplet-based, Hybrid Sparse-Dense Accelerator for Personalized Recommendations

 Transformers (2)

SpAtten: Efficient Sparse Attention Architecture with Cascade Token and Head Pruning
Switch Transformers: Scaling to Trillion Parameter Models with Simple and Efficient Sparsity

 VLSI Design (1)

Analog VLSI implementation of Support Vector Machine Learning and Classification

 Wireless Communications (1)

Massive MIMO Channel Estimation using Deep Neural Networks

 course assignment (3)

Minimum Edit Distance
UV Decomposition
Singular Value Decomposition

 data mining (3)

Minimum Edit Distance
UV Decomposition
Singular Value Decomposition

 hackathon (1)

Smart India Hackathon - 2019

 hardware (1)

Smart India Hackathon - 2019

 machine learning (1)

Word Predictor Using RNN

 numpy (1)

Word Predictor Using RNN