Description:

Large language models (LLMs) are powerful tools that can be used for a variety of tasks, such as generating text, translating languages, and writing different kinds of creative content. However, LLMs can be expensive to train.

In this talk, I will discuss how to train your own LLMs on a limited budget. I will cover topics such as data selection, model architecture, and training optimization. I will also provide a hands-on demo of how to train an LLM using a free cloud service.

This talk is for anyone who is interested in learning more about LLMs and how to use them. So come and learn how to unleash the power of LLMs on your own projects!

Here are some of the key points that I will cover in my talk:

• How to select the right data for training an LLM
• How to design an effective model architecture
• How to optimize the training process
• How to use a free cloud service to train an LLM I will also provide a hands-on demo of how to train an LLM using a free cloud service. This will give you a chance to see how the process works and to ask any questions that you have.

I hope that you will find this talk to be informative and exciting. I believe that LLMs have the potential to revolutionise the way that we interact with computers. By learning how to train your own LLMs, you can be a part of this revolution.

Prerequisites:

• Basic knowledge of large language models (LLMs)
• Some experience with programming
• A willingness to learn new things/Curiosity

Speaker Info:

Rahul Bhatia is currently working at Fidelity Investments as a Part of the Personal Investing group, AI Center of Excellence, Fidelity Investments. His focus revolves around working on building cutting edge Data Science capabilities to drive business success. Some of the focus areas include Natural Language Understanding, Abstractive Language Generation, Conversational AI, Few Shot learning, Reinforcement Learning. Rahul caters to multiple AI initiatives in the Large Language model space in the company and has been spending a lot of time on the cutting edge research and implementing these solutions to solve business use-cases. Rahul has spoken at multiple conferences prior to this, including PyCon Indonesia and PyCon Malaysia on topics such as Advanced Feature Engineering, Interpretable Machine Learning. Rahul has more than 3 years of professional experience working with both Large Organisations and Startups, focused on deriving useful business insights from large datasets of all kinds, and helping organisations generate business value using a data-driven approach.

To know more about Rahul, feel free to connect on LinkedIn - https://www.linkedin.com/in/rbhatia46/

Speaker Links:

https://github.com/rbhatia46

https://www.linkedin.com/in/rbhatia46/