GithubHelp home page GithubHelp logo

qunbb / deeplearning Goto Github PK

View Code? Open in Web Editor NEW
216.0 1.0 38.0 1.93 MB

All about DeepLearning: 推荐系统、自然语言处理、Tensorflow、Pytorch等

Python 51.03% Jupyter Notebook 45.38% Makefile 0.36% Starlark 0.67% C++ 2.56%

deeplearning's Introduction

1. tensorflow使用

项目文件夹:_tensorflow

  • batch_normalization: 专栏
  • dataset.shuffle、batch、repeat: 专栏
  • tfrecord读写数据: 专栏
  • estimator模型训练: 专栏
  • tf版本hashtable专栏

1.1 部署

项目文件夹:_tensorflow/serving

  • tensorflow serving: 专栏

1.2 自定义算子

项目文件夹:_tensorflow/tensorflow-custom-op独立实现git仓库

  • 如何实现TensorFlow自定义算子?: 专栏

2. 多任务学习MTL

项目文件夹:multitasklearning

  • shared_bottom、mmoe、ple模型介绍: 专栏
  • 多目标优化-Uncertainty Weight、GradNorm、Dynamic Weight Average、Pareto-Eficient专栏

3. 推荐系统

项目文件夹:recommendation

  • ctr训练提速(超大batch size)-CowClip专栏

3.1 Match(召回)

项目文件夹:recommendation/match

  • 多兴趣召回MIND: 专栏

  • 多兴趣召回ComiRec: 专栏

  • 深入浅出地理解Youtube DNN推荐模型: 专栏

  • 引入对偶增强向量的双塔召回模型: 专栏

3.2 Rank(排序)

项目文件夹:recommendation/rank

  • ctr特征重要性建模:FiBiNet&FiBiNet++模型专栏
  • ctr预估之FMs系列:FM/FFM/FwFM/FEFM专栏
  • ctr预估之DNN系列模型:FNN/PNN/DeepCrossing专栏
  • ctr预估之Wide&Deep系列模型:DeepFM/DCN专栏
  • ctr预估之Wide&Deep系列(下):NFM/xDeepFM专栏
  • CTR特征建模:ContextNet & MaskNet(Twitter在用的排序模型)专栏
  • CTR之行为序列建模用户兴趣:DIN专栏
  • CTR之行为序列建模用户兴趣:DIEN专栏
  • CTR之Session行为序列建模用户兴趣:DSIN专栏

4. TensorRT & Triton

TensorRT:一种深度学习框架,提升GPU模型推理性能

Triton:TensorRT对应的模型服务化,实现模型统一管理和部署

4.1 Triton

项目文件夹:triton

  • TensorRT&Triton学习笔记(一): triton和模型部署+client:专栏

5. 自然语言处理NLP

项目文件夹:nlp

5.1 bert句向量

5.2 bert系列

BERT预训练代码(tensorflow和torch版本):NLP/masked_language_model独立实现git仓库

  • BERT模型系列大全解读专栏

6. 深度学习trick

项目文件夹:trick

带pt后缀的为pytorch实现版本,不带后缀的则为tensorflow版本。

  • 变量初始化(initialization)、分层学习率(hierarchical_lr)、梯度累积(gradient_accumulation): 专栏
  • Stochastic Weight Averaging (SWA)、Exponential Moving Average(EMA)专栏
  • (unbalance)分类模型-类别不均衡问题之loss设计 & Label Smoothing专栏

7. 多模态

项目文件夹:multimodal

7.1 Stable Diffusion

项目文件夹:Stable Diffusion

  • AI绘画Stable Diffusion原理之VQGANs/隐空间/Autoencoder专栏
  • AI绘画Stable Diffusion原理之扩散模型DDPM专栏

8. Embedding

项目文件夹:embedding

  • Embedding压缩之hash embedding专栏
  • Embedding压缩之基于二进制码的Hash Embedding专栏

9. 大语言模型(LLMs)

项目文件夹:llms

9.1 LangChain

LangChain开发入门教程::llms/langchain_tutorial

  • Model I/O(prompts、llms、chat model、output parsers)专栏
  • RAG/Retrieval(文档加载器、文本分割器、Embedding、向量数据库、检索)专栏
  • Tools/Agents(工具、function call、agent)专栏

deeplearning's People

Contributors

qunbb avatar

Stargazers

 avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar

Watchers

 avatar

deeplearning's Issues

something looks wrong in MMOE.py (pytorch version)

result of the main(I only changed the size of the input)
I create the test data of (4,8), the output was the same,

task_output tensor([[-0.0373, -0.0265],
[-0.0373, -0.0265],
[-0.0373, -0.0265],
[-0.0373, -0.0265]], grad_fn=)
task_output tensor([[-0.0157, 0.0096],
[-0.0157, 0.0096],
[-0.0157, 0.0096],
[-0.0157, 0.0096]], grad_fn=)
OrderedDict([('click', tensor([[0.4973, 0.5027],
[0.4973, 0.5027],
[0.4973, 0.5027],
[0.4973, 0.5027]], grad_fn=)), ('like', tensor([[0.4937, 0.5063],
[0.4937, 0.5063],
[0.4937, 0.5063],
[0.4937, 0.5063]], grad_fn=))])
click torch.Size([4, 2])
like torch.Size([4, 2])

Can you help me figure out this? Thanks

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.