Complete Generative AI for upskilling & learning initiatives

Explore how to set up a Python development environment in VS Code, create Conda environments, manage Python 3.12, run Jupyter notebooks, and install ipykernel.

Explore Python basics, including syntax and semantics, case sensitivity, indentation, comments (single and multi-line), variable assignment, type inference, and common syntax errors with practical code examples.

Learn how to declare and assign variables in python, including naming conventions and common pitfalls. Grasp variable types, dynamic typing, type checking and conversion, with practical examples like a calculator.

Explore basic data types in Python, including integers, floating point numbers, strings, and booleans, and practice type conversion and handling common errors.

Explore python operators by implementing arithmetic, comparison, and logical operations with practical examples, building a simple calculator that prints results for addition, subtraction, multiplication, division, flow division, modulus, and exponentiation.

Explore Python control flow through conditional statements, if, elif, and else, with practical examples, nested conditions, common errors, and real-world use cases for upskilling and learning initiatives.

Explore lists in Python within a complete generative AI upskilling course, empowering learning initiatives with practical guidance on handling Python lists.

Explore tuples in Python: create and access immutable ordered collections, pack and unpack values, slice and concatenate, and work with count, index, and nested tuples.

Explore dictionaries in Python to map keys to values, enabling efficient data lookup and manipulation within upskilling initiatives powered by generative AI.

Explore real world uses of Python lists across to do lists, grades, inventory, and feedback, using append, remove, and membership checks to manage tasks and analyze data.

Explore functions in Python: define and call functions, use parameters (positional, keyword, default), handle variable-length arguments, and return values for reusable, readable code.

Learn to create anonymous Python functions with the lambda keyword, using multiple arguments and a single expression. See how map applies lambdas to square lists and filter usage.

Learn how the map function in Python applies a function to every item in an iterable, returning a map object; explore using lambda, multiple iterables, and practical examples.

Discover how the Python filter function creates an iterator by applying a condition to an iterable, filtering lists or dictionaries with even checks, lambda functions, and multiple criteria.

Learn to import and use Python modules and packages, from built-ins like math to external libraries like NumPy. Master import styles, aliases, and basic package creation to organize code.

Learn Python file operations for text and binary files: read entire content, line by line with newline handling, write and append, use read-write modes, and copy content between files.

Explore working with file paths using the os module: create directories, list files, join paths, check existence, and distinguish absolute, relative, and current working directories.

Master Python exception handling with try, except, else, and finally blocks to gracefully manage errors and keep program execution running.

Explore object oriented programming with classes and objects as blueprints, and learn to define attributes and methods with __init__ and self through a bank account example.

Explore Python inheritance through single and multiple inheritance, using parent and child classes, the super and __init__ methods, and real-world examples like car and Tesla.

Master polymorphism in Python through method overriding and abstract base classes, enabling objects from different classes to share a common interface and vary behavior.

Explore encapsulation and abstraction in Python by wrapping data and methods, hiding complex details, and using getter and setter methods to safely manage public, protected, and private variables.

Explore abstraction in oops by defining abstract base classes and abstract methods in Python, showing how hiding the complex implementation details lets child classes implement required features.

Explore Python magic methods, or dunder methods, and implement __init__, __str__, and __repr__ to customize object behavior for arithmetic, comparisons, and string representations.

Master operator overloading in Python by implementing magic methods to customize operations for a vector class, including addition, subtraction, multiplication, comparisons, and complex numbers.

Learn to build interactive ML and data science web apps with Streamlit, the open source Python framework. Install, import, create widgets, display data frames and charts, and run apps.

Create an end-to-end ml app with Streamlet that loads the iris dataset, trains a random forest classifier, and predicts species using interactive sliders.

Learn how to build an NLP roadmap from Python text preprocessing to vectorization, covering tokenization, bag of words, tf-idf, word embeddings, RNN, LSTM, GRU, transformers, and spam classification.

Learn text preprocessing with NLTK, focusing on stemming techniques such as Porter Stemmer, Regas Stemmer, and Snowball Stemmer, and compare them with lemmatization for natural language processing.

Learn to implement bag of words in language processing with NLTK, including stop words removal, stemming, and using count vectorizer with max features and binary options on a spam dataset.

Explore NLP in deep learning and sequential data, from ANN and CNN basics to RNN variants, encoders and decoders, self-attention, transformers, and generative AI with LLMs.

Apply z score to standardize features and compare scores across distributions, using mean and standard deviation, with Gaussian distributions and sklearn's standard scaler.

Explains forward propagation in a simple RNN, using one-hot word vectors and a three-neuron hidden layer with a self loop, and how time steps carry context through weights and biases.

Transform features with sklearn: drop columns, encode geography and gender with one hot and label encoders, scale with standard scalar, and save preprocessors with pickle for artificial neural network training.

Explore step-by-step training of an artificial neural network, covering optimizers and loss functions to enhance generative ai upskilling initiatives.

Load a trained ANN model from pickle files, encode geography with one-hot encoding and gender with label encoding, scale inputs, and predict customer churn.

Integrate an artificial neural network with a Streamlit web app to deploy a customer churn predictor, loading the trained model, encoders, and scaler, and providing interactive inputs for real-time predictions.

The lecture guides implementing a regression artificial neural network for a churn modeling dataset with salary as the target, covering preprocessing, encoding, scaling, training with early stopping, and evaluation.

Learn to determine the optimal hidden layers and neurons for an ANN using grid search, cross-validation, and heuristics, with a Keras classifier to identify the best model.

Develop an end-to-end sentiment analysis project using a simple RNN on the IMDB dataset, including data preprocessing, embedding layer, model training, and deployment with a Streamlit app.

Learn how embedding layers convert words into dense vectors for neural networks, replacing costly one-hot encoding, with word2vec guidance and 300-dimension feature representations for RNN inputs.

This lecture demonstrates a word embedding workflow in Keras TensorFlow, converting words to indices, padding sequences, and using an embedding layer to generate 10-dimensional vectors for a 10,000 word vocabulary.

Learn to build a text classification model using a simple RNN on the IMDB dataset, including loading data, embedding, sequence padding, and preparing train/test sets.

Build an RNN with an embedding layer that converts words to 128-d vectors. Train with a dense sigmoid output using binary cross-entropy, Adam, and early stopping.

Train a simple rnn on IMDB data to 94% and 81% accuracy, save the model as h5 file, and build a sentiment prediction pipeline with preprocessing and streamlet app.

Build an end-to-end Streamlit web app that loads a trained RNN model, preprocesses text, and predicts sentiment from IMDb reviews, with deployment to Streamlit cloud.

Explore the LSTM RNN architecture with forget gate, input and candidate memory, and output gate, plus the memory cell linking long-term and short-term memory via x_t and h_{t-1}.

Explore how the input gate and candidate memory in LSTM RNNs control information flow by combining x_t and h_{t-1} through sigmoid and tanh to update the cell state.

Explore how LSTM RNN trains text sequences with forget, input, and output gates and memory cells. See how word2vec embeddings and memory mechanisms predict restaurant reviews as good or bad.

Explore variants of LSTM RNN, including peephole connections and coupled forget-input gates, and compare to GRU, detailing gates, memory cell dynamics, and training implications.

Develop an end-to-end next word prediction model using lstm and gru. Learn data collection from Shakespeare Hamlet, tokenization, padding, pickle saving, embedding, early stopping, and streamlit deployment for real-time predictions.

Explore predicting next word with an LSTM model by tokenizing input, padding sequences, and predicting via model.predict, then save the model and tokenizer for an end-to-end streamlit app.

Build an end-to-end streamlet web app that loads a trained LSTM model and tokenizer to predict the next word from a text sequence.

Learn how bidirectional RNN and bidirectional LSTM RNN harness forward and backward context to improve sequence predictions, including embedding, and coverage of one-to-many, many-to-one, and many-to-many tasks.