[AI] 通用智能 - Agent

08/24

过去两年，行业迎来巨变。游戏规则不一样了，需要博客整理下思路。

除了下面的提到的几个技能点，Kubenates需要精进 正好结合近期的实际项目。

• [Full-stack] 一切皆在云上 - AWS

 

 

 

数字人

LivePortraint的出现，在工程的设计与实现上颠覆了许多。

[GenerativeAI] Avatar solutions

[Avatar] Avatar by ImgGen

[Avatar] Talking Face Dataset and Solutions

Animate Anyone

Face2Faceρ : Real-Time High-Resolution One-Shot Face Reenactment

[facefusion] Next generation face swapper and enhancer.

• • FaceFusion 源代码不是很多但实用，可以学习

 

 

Practice

[GenerativeAI] Webui Practice

LLM Fine Tune 

LLM Deployment

[MLLM] MiniGPT-4

 

 

发展历程

[GenerativeAI] Contrasive Learning and CLIP

[Multimodal] Since ViLT

[GenerativeAI] How to train both?

[GenerativeAI] GAN to Diffusion

[GenerativeAI] MAE and SAM

[diffusion] Stable Diffusion

[diffusion] Video Diffusion

 

 

Agent 

AI TREND JAN 2024

• • Lepton AI, 千行demo源代码，有必要研究

LangChain Tutorial

Agent Chatting Group [仿真智能聊天模拟器]

[agentic] Why AI Agent?

 

 

CS 194/294-196 (LLM Agents) - Lecture 1, Denny Zhou

A few examples to learn Reasoning Process?

中间过程的重要性。

But how to remove these in the prompt: let's think step by step.

 

[Submitted on 15 Feb 2024 (v1), last revised 23 May 2024 (this version, v2)]

Chain-of-Thought Reasoning Without Prompting

Xuezhi Wang, Denny Zhou

 

Ask LLM to self-select the most consistent answer.

无关信息作为“噪音”，语序描述的顺序改变。

 

 

CS 194/294-196 (LLM Agents) - Lecture 2, Shunyu Yao