Giter Site home page Giter Site logo

jkhumor / layoutocr_magic_llava Goto Github PK

View Code? Open in Web Editor NEW

This project forked from valeriawong/layoutocr_magic_llava

0.0 0.0 0.0 13.68 MB

License: GNU Affero General Public License v3.0

Shell 7.18% JavaScript 2.16% Python 88.26% CSS 0.39% HTML 1.66% Dockerfile 0.34%

layoutocr_magic_llava's Introduction

layoutocr_magic_llava

一个可以发论文的idea,欢迎原来做nlp或者cv的佬进行尝试,该过程能够学习实践大模型的pt及sft(助手君说a100单卡算力给够,分布式训练尝试中) 如有兴趣欢迎戳飞书

从0到1复现llava架构(可能类似llavar,不过会针对视觉模型语言模型进行替换提升layout ocr能力)有可能的任务包括:

  • 修改/微调/替换视觉模型以提升目标检测能力,
  • 替换llm为internlm2等 进行两阶段的训练:
  • 第一阶段,冻结视觉模型和文本模型,使用图像-文本对训练adapter;
    • 线性/非线性/自定义网络
  • 第二阶段,仅冻结视觉模型,使用图像-文本对端到端微调下游任务
    • 下游任务可增加其他类型

近期任务

  • 分布式训练尝试(单卡+单卡,单卡+受限卡)
  • 阅读相关论文
  • xtuner跑一下已有的internlm_llava_pt
  • llava更换vicuna为internlm2(或其他模型)进行两阶段训练(Obj:跑通流程,熟悉数据集格式)
  • 使用layoutlmv3、cogvlm等数据集训练vit
  • 垂直领域微调LLM
  • 数据增强和合成
  • 针对已魔改的视觉模型及文本模型进行两阶段训练
  • RAG
  • AGENT

后期可能还会实现

  • 尝试更改adapter的架构,增强连接表示学习
  • 借鉴AE的**进行数据增强与合成
  • RLHF

layoutocr_magic_llava's People

Contributors

haotian-liu avatar chunyuanli avatar valeriawong avatar mattmazzola avatar eggry avatar yorickvp avatar eltociear avatar didier-durand avatar guanlaoda avatar lengxiaom0 avatar mao-code avatar van-gogh avatar hill2hill avatar abdullamatar avatar abdul avatar timabdulla avatar dribnet avatar tonywang10101 avatar fvaysh avatar winglian avatar hirethehero avatar yvrjsharma avatar zhaoyangli-nju avatar paradoxzw avatar heltrix avatar kishida avatar l-salewski avatar omahs avatar payne911 avatar philokey avatar

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.