Code Monkey home page Code Monkey logo

Hey, I am yuyouyu! ✨

💬 About me

  • 🎓   Currently pursuing a Ph.D. (straight Ph.D.) degree in Computer Science at Shanghai University.
  • 🤔   Learning AI technology, git focusing on Python and Rust.
  • 🌱   Learning cutting-edge large model-related techniques, LLM, SD, and multimodal models (trying to catch up with developments)
  • 🎮   Enjoys gaming, including playing and making games, with a passion for incorporating cutting-edge AI technology into games.

🏢 Work Experience

  • 腾讯   📌 2024-03 —— now

    • 工作岗位:算法实习生 光子技术研发部
    • 工作内容:学习和研究LLMs的训练和部署,并应用于游戏角色扮演,撰写相关论文与技术报告;支持游戏相关舆情分析与周报生成。
  • 莉莉丝游戏   📌 2023-06 —— 2024-02

    • 工作岗位:算法实习生 研发中台
    • 工作内容:负责研究和优化主流AIGC技术在游戏生产中的应用,包括绘画、文本和动画生成,并协助技术落地和内容生态的发展。(主要是NLP LLM方向)
  • 上海卓越睿新数码科技股份有限公司(智慧树)   📌 2023-04 —— 2023-06

    • 工作岗位:大模型算法实习生 算法部门
    • 工作内容:利用高等教育领域的数据,搭建大模型的评价体系和fine-tuning数据集用于训练各学科的领域模型,搭建教育场景NLP大模型
  • 格物钛(上海)智能科技有限公司   📌 2021-05 —— 2021-09

    • 工作岗位:算法实习生 算法开发部门
    • 工作内容:负责机器学习数据管理平台的Python SDK和CLI开发,包括设计、测试、文档编写,并优化自动化深度学习模型迭代流程,同时研究并应用前沿算法理论于业务场景。

📅 Recent Projects

🔥 AFK(剑与远征) 虚拟玩家   📌 2023-11 —— now

  • SLG模拟玩家实现,在AFK(剑与远征)游戏中模拟真实玩家,根据玩家历史数据构建玩家行为模型,利用Lora微调LLM模仿玩家对话,给开源模型注入游戏知识,以提高在新服和人数较少的服务器中的玩家互动体验。

🤖 基于Agent驱动的故事推演与智能交互   📌 2023-10 —— 2023-11

  • 控制世界的LLM的模型会根据玩家,agent和当前世界状态进行故事的推演;控制Agent的LLM则会基于Agent自身状态,和玩家的关系,以及故事进行对话。

💎 Dislyte(神觉者) 虚拟角色   📌 2023-09 —— 2023-10

  • 利用RAG,LLM,Vector DB等技术制作Dislyte陪聊机器人,实现根据游戏世界观,游戏角色设定来回答玩家问题,探索了具有长期记忆,和一定情感分析能力的陪伴型ChatBot框架。

📡 模仿学习对战AI Bot   📌 2023-07 —— 2023-10

  • 模仿学习实现AI对战Bot,应用于Dislyt和Farlight 84(Google Play 2023年度最佳多人对战游戏)游戏中,在Dislyte中模仿学习AI对比规则AI拟人度提升20%左右,在Farlight 84中模仿学习AI训练成功3种风格AI,并针对不同段位层级实现了AI的有效投放。
c3quy-ybyv5.mp4
2rnch-pp4o4.mp4

😺 Samo(万龙觉醒)立绘魔宠生成   📌 2023-06 —— 2023-08

  • 根据Samo项目立绘风格训练SD模型生成魔宠图片,玩家上传宠物照片来生成Samo游戏中风格的魔宠,总PV次数为117K次,总UV人数为50K,且出现明显自传播效应,分享点击率高达55%,后台追踪用户总生图次数为41.8K次。

👩🏻‍🏫 教育大模型研发   📌 2023-04 —— 2023-06

- - 智慧树**网页“大明白”AI助手和“知到-教师版”APP的算法模型开发**,涉及到成本控制。调研主流大模型如LLaMA、chatGLM、MOSS的性能,包括错别字识别等任务,并使用内部教育数据集对不同学科模型进行Fine-tuning。

📊 Github metrics



Ashutosh's github activity graph

yuyouyu's Projects

aaeg icon aaeg

Paper title: Adversarial Domain Adaptive Embedding Generative Network: Transfer learning between different properties of HEAs

baiducitiaospider icon baiducitiaospider

BaiduCitiaoSpider is a web scraper designed to crawl and extract information from Baidu Baike

library icon library

上海大学数据库二大作业,图书管理系统

mcvn icon mcvn

Material composition visualization network for materials properties prediction

port-scanner icon port-scanner

Shanghai University 2021 Information Security Course Port Scanning Program

preparatory_materials icon preparatory_materials

Implementation Plan for the 23rd Preparatory Party Member Training at the School of Computer Science of Shanghai University Party School

reinforcematdesign icon reinforcematdesign

A pioneering project leveraging Reinforcement Learning for inverse material design, enhanced with Large Language Models for efficient process filtering.

vim icon vim

My vim config files.

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google ❤️ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.