 ## Neural Network Module

In this assignment, we have covered, use cases of tensorflow in real world, tensors definition, tensor types, tensor formats, tensorflow methods, mathematical operations in tensorflow, numpy compatibility, gradient and finally a small end to end linear classifier using the concepts of tensorflow we learnt from the beginning.

In this this assignments we have mentioned every fundamentals of Pytorch which one should know while working on this library.

In this assignment we have mentioned all the concepts that one should know before getting into neural networks. Each and every concept that is required for working on neural nets has been mentioned in this assignment.

In this assignment, we have covered definition of activation function, its uses in deep learning, types of activation function like linear, binary & non linear and their subtypes. We have also covered how to use it in coding.

In this assignment we have trained multiple models using mnist handwritten digit dataset and learnt how to choose number of layers and number of neurons for the model. We have also explained how changing the architecture of the neural network affects the model.

In this assignment, we will see what is a loss function, how to choose a loss function for a particular problem. We have covered multiple cases like regression, binary and multi-class classification where you will choose loss function accordingly.

In this assignment we will learn how we go from left to right in a neural network via forward propagation and then we do back propagation to update the weights in our neural network which eventually helps us in reducing our loss.

In this assignment, we have covered what is an optimizer, its different types like gradient descent, sgd, momentum, adagrad, adadelta, adam, mini batch, RMSprop, NAG,  FTRL  with a working example .

In this assignment, you will learn different types of regularization techniques like dropout, L1, L2 and early stopping using the example of fashion mnist data. You will build multiple models to get deeper insights.

This assignment will teach you to code neural network completely from scratch. After this assignment one will be very much comfortable on creating any kind of neural nets from scratch and that’s what the main agenda of this course.

In this assignment, you will implement all learnings from the neural network assignments on multiple datasets.

## Computer Vision Module

In this assignment, we will cover basic concepts of CNN like what is CNN, padding, pooling, striding, fully connected layer using cifar dataset.

In this assignment, we will learn need of preprocessing, image augmentation, steps to use in the preprocessing, use of imagedatagenerator, and other steps using Tensorflow.

In this assignment you will learn difference between fit and fit generators, what is flow from directory method,  in the tensorflow by implementing it  on dog cat classification dataset.

In this assignment we will learn what is transfer learning, why do we need it, what is resnet, resnet architecture, how to implement transfer learning using resnet etc.

The main purpose of this assignment is to get a complete idea of Alexnet and code it completely from Scratch.

The main purpose of this assignment is to get a complete idea of VGG-16 and code it completely from Scratch.

In this assignment, we will learn functionalities of opencv. Opencv is a large open source library for computer vision, machine learning and image processing. We will learn how to process an image and video too.

In this assignment, we will first teach you installation of yolo and its dependencies, getting the data, data annotation, how to create bounding boxes , how to train different yolo models, tensorboard implementation, and making predictions.

In this assignment, we will learn how to use detectron for object detection, image segmentation and panoptic segmentation.

In this assignment we will learn what is image segmentation, it’s importance, types of image segmentation, for example region based segmentation, Edge Detection Segmentation.

In this assignment we will see how we can use DeOldify  to colorize black & white images.

In this assignment we will see what is GANs, it’s applications and how to implement it on the dataset.

## NLP Module

In this assignment, you will learn about NLP, top 8 use cases in industry, concepts of tokenization, punkt, lemmatization, stemming , stemming vs lemmatization and stop words removal. All these concepts will help you in dealing with sentences and paragraphs later.

This assignment covers concepts of  countvectorizer, bag of words, TD-IDF, N-Grams, POS tagging which will teach you how you can convert words into vectors. It will help you dealing with text data. You will find all the necessary resources to do the assignment inside the file.

This assignment covers advanced concepts like Gensim, Word2Vec, continuous bag of words(CBOW), skip-Grams which will teach you more ways of converting words into vectors. It is extended version of assignment 2 converting words into vectors.

It is a project on predicting sentiment of words using logistic regression technique and all the learnings we had in previous 3 assignments. It covers countvectorizer, lemmatization, stemming, TF-IDF etc.

In this assignment, we will revise basics of artificial neural network with practical implementation using the dataset of credit card churn modelling.csv.

In this assignment you will build and train a model using RNN on the mnist dataset. You will learn use cases of RNN by practical implementation.

In this assignment we will learn what Is LSTM, it’s implementation using RNN using imdb dataset.

In this assignment, we will learn about GRU, it’s architecture, applications, Bidirectional LSTM RNN and the implementation of bidirectional LSTM using Tensorflow.

In this assignment, we will learn about encoders, decoders, hidden state, attention model transformer, positional encoding, residuals, hugging face, and transformers pipeline.

In this assignment we will see implementation of bert model in two steps, pretraining and fine-tuning.  We will do fine tuning for QA, sentiment analysis and Named Entity Recognition(NER).

The Autocorrect with NLP project provides to way to get recommendations of words when you start typing the word. It also recommends correct spellings for wrongly spelled words based on text distance. It has an interface which is created in Google Colab.

## Time Series Module

In this assignment, we will learn basics of time series i.e what is time series, its use cases with real world examples, time series components, and we will use a simple dataset to visualize time series data.

In this assignment, we will learn what is called stationarity series, how we check stationarity in the data using visualization, augmented dickey fuller test, KPSS test.

In this assignment, we will cover topics like correlation, autocorrelation, partial autocorrelation function and differencing using multiple datasets to explain each concepts.

In this assignment, we will learn about arima models, terms used in arima, AR & MA models, order of differencing, order of AR term, order of MA term, how to build arima model, and metrics used in it.

In this assignment, we will cover auto arima, interpreting residual plots in arima model, automatically building sarima model, and building sarimax with exogenous variable.

In the assignment, we will cover model selection, AIC, BIC, and it’s implementation using drug sales data

In this assignment, we will cover varma with its implementation.

In this assignment, we will cover auto regressive concepts with example, model and implementation.

In this assignment, we will see the implementation of ANN for time series forecasting on stock prices dataset.

In this assignment we will cover the effect of CNN on numerical data and implement CNN for time series forecasting.

In this assignment you will be learning about implementing time series forecasting using LSTM neural networks.

Scroll to Top