Deep dive into understanding GenAI model working

 GEN AI Generative AI is a type of artificial intelligence that uses machine learning to create new patterns to create new content based on patterns and relationships in existing data. Neural networks are used to learn from data, and then use those learned patterns to generate new content. Steps to create your own Gen AI model: Defining your goal Data collection Data preparation and preprocessing Choose the framework you would like to work with(e.g. Tensorflow, PyTorch, Keras) Choose the model architecture(e.g. Transformers or GAN) Training the model Evaluate and refine Objective: To develop an end-to-end gen AI model for medical summarization. Input: MRI images of the brain Output: Summarized medical report  Here we already have a deep learning model based on MRI images of the brain trained on Parkinson's and Alzheimer's disease. To develop a gen AI model based on these we have to follow these steps: Choosing the text summarization method. There are two approach  Extrac...

                                Gemma model learning Gemma is an open-source model created by Google AI. It is a type of large language model. It's designed to be a versatile building block for various tasks related to natural language processing (NLP). Objective: Develop end to end product to suggest gift products Prompt: "I want to buy a 12-year-old boy a gift who likes outdoor games" For this prompt generate gift suggestions Steps followed: Loading all the necessary libraries and running the model Enter the prompt for the model Generate the output The generated output is: Prompt: Here are some great outdoor gift ideas for a 12-year-old boy who enjoys being active: Output: Single Prompt Result: Here are some great outdoor gift ideas for a 12-year-old boy who enjoys being active: **Outdoor Gear:** * **Backpack:** A high-quality, durable backpack that can handle the weight of a full day of hiking, cam...

                   Hugging face's learning Hugging Face is a French-American company with a focus on building tools and resources for developers and researchers working in the field of artificial intelligence, particularly natural language processing (NLP). Image-to-text models Pix2struct   1. Pix2Struct is a pre-trained image-to-text model developed by Google AI. It's designed              for tasks involving visually-situated language understanding.     2.   Purpose: Understand and interpret the structure of web pages based on screenshots. Bridge the gap between visual information and textual understanding.     3. Functionality: Takes a masked screenshot of a webpage as input (where parts of the image are hidden). Predicts the underlying HTML structure corresponding to the visible parts of the screenshot. This process essentially translates visual elements like text, butt...