Search result for:
On-demand webinar
Speakers
Ines Chami
Co-founder & Chief Scientist
Numbers Station
Ines Chami is the Chief Scientist and Co-Founder of Numbers Station. She received her Ph.D. in the Institute for Computational and Mathematical Engineering from Stanford University where she was advised by Prof. Christopher Ré. Prior to attending Stanford, she studied Mathematics and Computer Science at Ecole Centrale Paris. Ines is particularly excited about building intelligent models to automate data-intensive work. Her work spans applications such as knowledge graph construction and data cleaning. For her work on graph representation learning, she won the 2021 Stanford Gene Golub Doctoral Dissertation Award. During her Ph.D. she interned at Microsoft AI and Research and Google Research where she co-authored the graph representation learning chapter of Kevin Murphy's "Probabilistic Machine Learning: An Introduction" book.
Shane Johnson
Senior Director of Product Marketing
Snorkel AI
I started out as a developer and architect before pivoting to product/marketing. I'm still a developer at heart (and love coding for fun), but I love advocating for innovative products -- particularly to developers.
I've spent most of my time in the database space, but lately I've been going down the LLM rabbit hole.
Combining structured and unstructured data to enhance RAG capabilities
The reasoning capabilities of large, state-of-the-art open models such as Meta’s Llama 3.1 405B rival those of popular closed models, but the compute resources required for low-latency inference often put them out of reach.
However, LLM distillation can be used to create small language models (SLMs) for specialized tasks that preserve the reasoning capabilities of large models while significantly reducing inference costs.
In this webinar, we’ll provide an overview of LLM distillation, explain how it compares with fine-tuning, and introduce the latest techniques for training SLMs using larger models and knowledge transfer.
Watch this webinar on-demand and learn how to:
- Train small-language models (SLMs) for specialized tasks
- Choose between LLM fine-tuning and distillation
- Reduce inference costs while preserving response quality