How to train your own Large Multimodal Model — with Hugo Laurençon & Leo Tronchon of HuggingFace M4
Latent Space is heating up! Our paper club ran into >99 person Discord limits, oops. We are also introducing 2 new online meetups: LLM Paper Club Asia for Asia timezone (led by Ivan), and AI in Action: hands-on application of AI (led by KBall). To be n...
Full show notes and links
No “Zero-Shot” Without Exponential Data: Pretraining Concept Frequency Determines Multimodal Model Performance
Multimodal models' "zero-shot" performance relies on pretraining data; exponential data increase is needed for linear improvements, challenging true "zero-shot" generalization.
https://arxiv.org/abs//2404.04125
YouTube: https://www.youtube.com/@ArxivPapers
TikTok: https://www.tiktok.co…
[short] No “Zero-Shot” Without Exponential Data: Pretraining Concept Frequency Determines Multimodal Model Performance
Multimodal models' "zero-shot" performance relies on pretraining data; exponential data increase is needed for linear improvements, challenging true "zero-shot" generalization.
https://arxiv.org/abs//2404.04125
YouTube: https://www.youtube.com/@ArxivPapers
TikTok: https://www.tiktok.co…
[QA] No “Zero-Shot” Without Exponential Data: Pretraining Concept Frequency Determines Multimodal Model Performance
Multimodal models' "zero-shot" performance relies on pretraining data; exponential data increase is needed for linear improvements, challenging true "zero-shot" generalization.
https://arxiv.org/abs//2404.04125
YouTube: https://www.youtube.com/@ArxivPapers
TikTok: https://www.tiktok.co…
Simon Tokumine - Gemini and the future of Generative AI tools
Meet Simon Tokumine, a Director of Product Management at Google leading teams for large language and multimodal model AI tools. On this episode of People of AI, hosts Ashely Oldacre, Luiz Gustavo Martins, and guest Simon chat about Gemini, Google's latest AI multimodal model. Join us as we discuss …
What Comes After ChatGPT? A Preview of the Future of LLMs
Today on The AI Breakdown, NLW describes three bits of news from this week that together help us get a preview of the future of large language models, including interpretability research from OpenAI using GPT-4 to better understand how GPT-2 works; Anthropic's "Constitutional AI" principles; and Me…
Wed. 09/11 – Mistral Goes Multimodel; And An AI Model Mystery
Mistral goes multimodal for the first time. Meta admits to scraping the data of every adult Australian. The details on the new PS5 Pro. Wouldn’t it be wild if, through stablecoins, crypto BECOMES the banking system instead of replacing it. And a weird mystery in AI land.
Sponsors:
Promevo.com/tech…
7 Use Cases for GPT-4o
Explore seven innovative uses for OpenAI’s new GPT-4o model, a natively multimodal model capable of handling text, audio, and visual inputs. Discover how these capabilities can be applied in various fields like marketing, customer service, tutoring, and more.
**
Check out the hit podcast from HBS M…
Insights from a Kaggle Grandmaster: Multimodal Models, Agents, Document AI & more | Leading with Data Ep 49
This week, we’re joined by Mark Landry, Director of Data Science & Product at H2O.ai and a top-ranked Kaggle Grandmaster. With extensive experience supporting data science for various H2O.ai products and award-winning applications developed with PwC, Mark’s journey spans data science, business inte…
Generative AI and Multimodal Models — What We Can Expect in 2025
Video - https://youtu.be/J-2TBZ1zYuA
Step into a world where AI not only creates original content but also understands multiple types of data at once. In this forward-looking piece, we unveil how generative AI and multimodal models could soon redefine everything from everyday tasks to high-stakes i…
PODCAST SEARCH ENGINE
You can use our website to search 3,836,712 podcasts and 197,720,769 episodes by people, places, or topics.
An investment in knowledge pays the best interest, so I listen to podcasts.
‐ Benjamin "Llamacorn" Franklin
EDIT
Thank you for helping to keep the podcast database up to date.