Let's explore "Gemini", Google's new multimodal AI model family. We'll examine how Gemini advances the state-of-the-art in large-scale language modeling, image and audio processing, and video understanding, building upon foundational work in sequence models, neural networks, and machine learning systems. Gemini Ultra sets new records in 30 out of 32 benchmarks, including text and reasoning, image understanding, video understanding, and speech benchmarks. Notably, Gemini Ultra marks a milestone in human-expert performance on MMLU, a benchmark for knowledge and reasoning.
Gemini Blog Post: blog.google/technology/ai/goo...
Gemini Technical Report: storage.googleapis.com/deepmi...
AI Bootcamp (preview drops on Christmas): www.mlexpert.io/membership
Discord: / discord
Subscribe: bit.ly/venelin-subscribe
GitHub repository: github.com/curiousily/Get-Thi...
Join this channel to get access to the perks and support my work:
/ @venelin_valkov
00:00 - Intro
00:13 - AI Bootcamp on MLExpert.io
01:26 - Google Gemini Blog Post
03:08 - Performance Overview
07:13 - Training & Inference Hardware (TPUs)
07:48 - Technical Report
08:30 - Model Architecture
11:11 - Training Infra
14:58 - Training Dataset
17:58 - Detailed Evaluation
21:42 - Long Context Effectiveness
22:55 - Conclusion
#chatgpt #gpt4 #artificialintelligence #googleai #deepmind #chatbot #llm
Негізгі бет Google Gemini - Can it Beat ChatGPT/GPT4? | Technical Report - Model Architecture, Dataset, Training
Пікірлер: 1