
Priya Shastri introduces the course on large language models. She shares 18 years as a software engineer and a commitment to reskilling as essential for success in technology and AI.
Explore what large language models are, their place in deep learning, foundational models, and applications like text classification, natural language processing, and summarization.
Trace the history of large language models from Eliza and RNN to transformers and GPT, highlighting milestones, training principles, and the role of GPUs in scaling.
Explore the architecture of large language models, including dual reading and pattern matching functions, transformer encoder-decoders, and massive parameter scales, plus training, fine-tuning, and multimodal capabilities.
Transformers transform input sequences into outputs via encoders, decoders, self-attention, and positional encoding, with layered matrix multiplications and token weights guiding large language models.
Explore how transformers scale with parameters, the encoder–decoder structure, and how self-attention, positional encoding, and matrix multiplication generate model outputs for Bert, Llama, and ChatGPT.
This course will help you understand the basics of Large Language Models (LLMs) and Transformers. Why are LLMs now so popular? What are its characteristics? Why do companies have their own LLMs?
Why is Data so important? What is the value of specialized data and how are they used in the models?
Once you answer these questions the idea of the large language models will become clear. Large Language models are trained models that can provide an accurate answer to specific questions in the domain. The domain can be engineering, science, politics, geography, computer science, sports, stock market or any other field where there is enough data to train a model.
When data is viewed as a commodity then the value of the data increases. You can start predicting response to questions that do not have answers. Can you predict who can win the SuperBowl, or the NBA based on data obtained from previous years?
Are you becoming clairvoyant? There is science behind all this mechanism.
Hope you enjoy this course and learn to predict the future or train your models to provide accurate responses. Good luck in your journey of Large Language Models. There is more to come in this space.