Sarthak Garg's repositories
Severstal-Steel-Defect-Detection
Steel Defect Detection using U-Net. Optimising training and inference using Automatic Mixed Precision and TensorRT respectively.
TrainYourOwnYOLO
Train a state-of-the-art yolov3 object detector from scratch!
OTOMYCOSIS
Ear Disease Detection using Deep Learning
Accelerating-Inference-in-Tensorflow-using-TensorRT
TensorRT optimises any Deep Learning model by not only making it lightweight but also by accelerating its inference speed with an idea to extract every ounce of performance from the model, making it perfect to be deployed at the edge. This repository helps you convert any Deep Learning model from TensorFlow to TensorRT!
Automatic-Mixed-Precision
NVIDIA's Automatic Mixed Precision(AMP) is applied on a language model used for trigger word detection. Main aim is to find out if AMP helps the model to converge faster.
Basics-of-Deep-Learning
Deep Learning Architectures from scratch
Deep-Learning-Coursera
Deep Learning Specialization by Andrew Ng, deeplearning.ai.
hands-on-transfer-learning-with-python
Deep learning simplified by transferring prior learning using the Python deep learning ecosystem
LungCancerDetection
Use CNN to detect nodules in LIDC dataset.
models
Models and examples built with TensorFlow
neural-doodle
Turn your two-bit doodles into fine artworks with deep neural networks, generate seamless textures from photos, transfer style from one image to another, perform example-based upscaling, but wait... there's more! (An implementation of Semantic Style Transfer.)
Severstal-Steel-Defect-Detection-Kaggle
in this repository i will put some of my kernels that i used in steel defect detection kaggle competition,competition link : https://www.kaggle.com/c/severstal-steel-defect-detection
Wake-UP-word-detection
Wake-up-word(WUW)system is an emerging development in recent times. Voice interaction with systems have made life ease and aids in multi-tasking. Apple, Google, Microsoft, Amazon have developed a custom wake-word engine, which are addressed by words such as ‘Hey Siri’. ‘Ok Google’, ‘Cortana’, ‘Alexa’. Our project focuses initially only detection and response to a customized wake-up command. The wake-up command used is “GOLUMOLU”. A wake-up-word detection system search for specific word and reads the word, where it rejects all other words, phrases and sounds. WUW system needs only less memory space, low computational cost and high precision. Artificial Neural Networks(ANN) have reduced the complexity, computational time, latency, thus the efficiency of system has improved. Deep learning has improved the efficiency of automatic speech recognition(SR), where wake word detection is a subset of SR but unlike keyword spotting and voice recognition. A deep learning RNN model is used for the training of the network. RNN are specifically used in case of temporal sequence data and has the ability to process data of different length but of same dimension. For training a model, labelled dataset is needed. We generated three forms of data: golumolu, negative and background. Such that, the model learns circumspectly and attentively detects when specific word found. To start communication with system, the wake word should be delivered. The main task of WUW detection system is to detect the speech, to identify WUW words among spoken words, to check whether the word spoken in altering context.