Generative AI Mastery With 15+ Real Time Projects

Explore what generative AI is and how large language models transform unstructured data into new text, images, or videos, powering tools like ChatGPT, Google Gemini, and LLaMA.

Explore the complete end-to-end generative AI pipeline from data acquisition and preparation to deployment and monitoring, with hands-on techniques for data augmentation, preprocessing, feature engineering, and model evaluation.

Master data preprocessing and cleaning for generative AI using a 50k IMDb dataset, including lowercasing, HTML and URL removal, punctuation handling, stopwords, tokenization, and stemming or lemmatization.

Learn data representation and vectorization for generative ai, turning text and image data into numerical vectors with one-hot, bag-of-words, tf-idf, and word2vec techniques.

Master text classification by preprocessing and vectorizing IMDb reviews with bag of words or tf-idf, and training simple models like Gaussian NB or random forest in a Colab demo.

Explore how large language models use transformer architecture to understand and generate human language, including encoder and decoder roles, transfer learning, and prompt engineering.

Master the transformer architecture and attention is all you need, covering encoder and decoder layers, self-attention, multi-head attention, positional encoding, residual connections, and normalization for language translation.

Explore how ChatGPT is trained through generative pre-training, supervised fine tuning, and reinforcement learning from human feedback, turning a GPT model into an interactive application via the OpenAI API.

Discover Hugging Face as a platform for generative AI, NLP, and computer vision, with models, datasets, pipelines, and transformer library, plus tokenization and fine-tuning of open source large language models.

Master practical use of the Hugging Face transformers library by building pipelines for sentiment analysis, text classification, translation, text generation, and more, using GPU Colab runtimes and model IDs.

Explore tokenization in Hugging Face using auto tokenizer and a Bert-based multilingual sentiment model to transform text into vector representations, with attention masks and input IDs.

Fine-tune pre-trained models with Hugging Face on the IMDb sentiment analysis task, using datasets, tokenizers, and training workflows to save the model and tokenizer.

Generate a Hugging Face API key by navigating to settings, creating a new access token, selecting read permissions, and securely saving the token for future use.

Learn text summarization with Hugging Face by fine-tuning a transformer on the Samsung dataset in Colab using a GPU, covering tokenization, data collators, training, and rouge evaluation.

Master text-to-image generation with diffusion models in Hugging Face, using the diffusers library and multi-modal prompts to produce vivid images.

Learn to generate text-to-speech with the hugging face transformers pipeline using the pseudo bark small model on cuda, and play the output audio via IPython display, following the Colab notebook.

Discover how OpenAI powers generative AI applications with GPT models, DALL-E, and whisper, including API keys, playground usage, and transitioning from free access to premium plans.

Generate your OpenAI API key, copy the secret key securely, and set read, write, and all permissions, then explore temperature, max tokens, and top P to control model output.

Set up your local environment with Anaconda Python interpreter, create environments, and install Visual Studio Code and Git with Git Bash to develop generative AI projects and explore OpenAI APIs.

Learn to use the OpenAI chat completion and completion APIs, set up the OpenAI package, manage API keys, and run hands-on Python examples for multi-turn prompts and single prompts.

Explore function calling in OpenAI to let a large language model interact with third-party APIs like weather data via RapidAPI, returning JSON responses and retrieving current weather by location.

Learn to implement a telegram chatbot using OpenAI’s gpt-3.5 chat completion model, connecting a telegram front end with the OpenAI back end via the io gram framework.

Fine-tune a GPT three classifier on custom data with prompt and completion pairs, using the OpenAI API to train in Colab for a baseball versus hockey task.

This project demonstrates audio transcript translation with Whisper, converting speech to text and translating to a target language using a GPT model, via a Flask web interface.

Learn to build an image generation app using DALL-E via the OpenAI API, sending prompts to generate images, with a Flask interface and demo prompts.

Master prompt engineering to structure prompts that coax useful outputs from large language models, integrating clear instructions, persona, and format while embracing zero-shot and few-shot prompting, and retrieval-augmented generation.

Explore vector databases, embeddings, and similarity search to store high-dimensional vectors for language models. Learn practical demos with Lang, Chroma DB, Pinecone, Neo4j, and Feis.

Learn to set up a local chroma db vector database, store embeddings, and run Colab demos that enable retrieval with LangChain for question answering.

Master Pinecone vector database to store embeddings in the cloud, create an index, set up API keys, install the Pinecone client, and perform similarity search with LangChain.

Master the cloud-based vector database V8 with free credits, credentials, and a Python workflow for embeddings, text splitting, storing vectors, and performing similarity search and question answering with Lang Chain.

Learn the LangChain framework for building generative AI applications, install and set up tools, access OpenAI and Hugging Face models, and run a practical Colab demo with memory and prompts.

Learn how to build robust prompts with Langchain prompt templates, using input variables like cuisine to avoid rewriting prompts, and set up templates for scalable apps.

Learn to pair a prompt template with a large language model using LangChain chains for multi-step workflows. Explore lm chain, simple sequential chain, and multi chain.

learn how Langchain agents and tools plug into Wikipedia and other plugins to fetch up-to-date data, and how to initialize tools and run an agent.

Explore memory in LangChain by implementing conversation buffer memory and conversation chain to give your AI apps context awareness and ChatGPT-like recall with an OpenAI LLM.

Learn to load documents with LangChain loaders for pdf, csv, json, html, markdown, and office formats, connecting Slack and Discord to extract data into a vector database for LLM use.

Explore multi dataframe agents in LangChain to work with pandas data frames using any large language model, such as OpenAI. Compare columns across three data frames.

Learn to use Hugging Face open source LLMs with LangChain, accessing models via API requests or downloading them in Google Colab, and build prompts for T5 and Falcon models.

Build a project called interview questions creator that uploads PDF documents to automatically generate interview questions and answers using a large language model, LangChain, and a vector database with embeddings.

Build a custom website chatbot using retrieval augmented generation with a vector database, LangChain, and embedding models to answer questions from website data via sitemap.xml and a knowledge base.

Explore open source large language models such as Llama 2, Falcon, and Google Palmetto, learn to run quantized models on Colab, and build end-to-end generative AI applications.

Learn to run open source LLMs with LangChain using Llama 2 variants on Hugging Face. Build chat and text generation pipelines with custom prompts, templates, and Colab GPU setup.

Build a custom website chatbot using open source llms by extracting site content with sitemap.xml, embedding with hugging face, indexing in pinecone, and answering queries with llama 2 via langchain.

Explore Falcon, an open source large language model, and learn inference with LangChain, Hugging Face, and embeddings to build end-to-end generative AI applications.