chainhead

Introduction

Installation

• https://www.keycloak.org/docs/latest/server_installation/
• https://medium.com/@hasnat.saeed/setup-keycloak-server-on-ubuntu-18-04-ed8c7c79a2d9
• Ubuntu 18.04, 4 GB RAM
• Pre-requisite: sudo timedatectl set-timezone Asia/Kolkata

Install Java

chainhead

Person:

• What is Generative AI?

ChatGPT:

• Generative AI refers to a subset of artificial intelligence (AI) techniques and models that are designed to generate new, original content or data that resembles human-created content. These models are capable of producing text, images, audio, and even other forms of data, often with a high degree of creativity and realism. Generative AI has made significant advancements in recent years, thanks in part to deep learning and neural network architectures.

One of the most prominent types of generative AI models is the Generative Adversarial Network (GAN), introduced by Ian Goodfellow and his colleagues in 2014. GANs consist of two neural networks, a generator and a discriminator, which work in opposition to each other. The generator tries to create content that is indistinguishable from real data, while the discriminator tries to differentiate between real and generated data. This adversarial process helps the generator improve its output over time.

Applications o

chainhead

Requirements

Summary

There are three Kafka topics t1, t2 and t3 where, t1 has raw data and t2 has look-up values. Message(s) in t1 are used to look-up t2. If a match is found, then the message t1 is enriched with values in t2 and written out to t3. Messages in t1 are expected at a rate of around 100 messages per second. The expected deliverable is a Quarkus project such that, the mvn command results in a working JAR file.

Details

Messages in t1 will be JSON as shown below. Each message will have a UUID as the key. This UUID is same as meta.id. The data array may have hundreds of entries.

chainhead

Notebook

Theory

LLM

• Build LLM from scratch in Python using createllm package - https://pythonscholar.com/build-a-large-language-model-from-scratch/
• Private LLM using databricks-dolly-15k (approximately 15,000 instruction/response fine-tuning records) - https://www.leewayhertz.com/build-private-llm/
• The Emergence Of Large Language Model (LLM) API Build Frameworks - https://cobusgreyling.medium.com/the-emergence-of-large-language-model-llm-api-build-frameworks-78d83d68eeda
• Corpus size and LLM - https://genai.stackexchange.com/q/613/2269