Weaviate Podcast Charles Packer on MemGPT - Weaviate Podcast #73!
6 snips
Nov 6, 2023 Charles Packer, lead author of MemGPT at UC Berkeley, discusses the concept of explicit memory management in GPT models, the use of prompts to handle memory limitations, interrupts in retrieval augmented generation (RAG), achieving ideal running speed in high parameter models, fine-tuning MemGBT for long conversations, search actions pagination, role-playing language models, and the future integration of memory in chatbot platforms.
Chapters
Transcript
Episode notes
1 2 3 4 5 6 7
Introduction
00:00 • 2min
Explicit Memory Management and Self-Awareness in GPT Models
02:17 • 4min
Exploration of Interrupts in RAG and Function Calling Standardization
06:38 • 9min
Ideal running speed of a meta model and working with billion parameter models
15:11 • 2min
Fine-tuning MemGBT for Long Conversations
17:26 • 18min
Search Actions Pagination and Role-Playing Language Models
35:13 • 11min
The Future of Memory in Chatbots
46:03 • 5min