Gemini, Google's Large Language Model
summary
Gemini is natively multimodal and ships as Nano, Pro and Ultra with a new Gemini 15 Flash variant that aims for faster, cheaper responses. It accepts up to a 1 million token context window (roughly 1 hour video, 11 hours audio, 30k lines of code or 70k words) and a 2M-token whitelist. Demos show Java integration via LangChain4j or the Gemini SDK, doing streaming responses, extracting POJOs from unstructured text, RAG with an in‑memory vector DB, and function calling to invoke APIs. There are also local small models Gemma 2B/7B.
AI-written summary · about
chapters (16)
- 0:00Introduction to session
- 1:53History of LLMs and Transformers
- 3:43Gemini model overview
- 8:43Gemini web app demo
- 16:28Gemini 15 Flash features
- 26:23Gemma open model introduction
- 28:51Using Java with Gemini
- 30:16Basic Java SDK examples
- 33:09Conversational AI with Java
- 35:17Multimodal interactions in Java
- 42:34Data extraction with LLMs
- 43:58Retrieval augmented generation
- 47:07Function calling with APIs
- 50:18Safety filters and limitations
- 51:22Training custom models
- 53:30Q&A and closing remarks
related talks
51mGemini, Google's Large Language Model
51mGemini, Google's Large Language Model
19mGemini 2.0: Beyond the Basics
16mOpen models in the Gemini era
56mBuilding Multimodal Apps with Gemini 2 0 and Gemini SDK
01h 01mBuilding Multimodal Apps with Gemini 2.0 and Gemini SDK
18mIntroduction to Gemini APIs and AI Studio
01h 16mI'm in my Gemini Era!