GithubHelp home page GithubHelp logo

semeval_2010_task8's Introduction

Remark for semeval 2010 task8

prework: 当前实现的最好效果为cnn + piecewise max pooling 83.6

10.13 在实现基准模型,cnn + max pooling + softmax 时发现,f1最好效果只有63 将cnn激活函数更换为relu,效果提升到71

10.15 上述结果有误,调整网络结构,发现max pooling实现有误,更换之后发现f1最好效果提升到82.58, 相差0.21 使用新版本的wiki词向量,发现此词向量的效果不如google news,f1效果为81.50 todo: 使用window=5版本的词向量试一下 todo: 为未登录词使用同一个词向量

10.16 开始建立新模型: TCA-CNN ??? ranking loss function 存在问题

10.17 调整了position embedding的输入部分,发现simple_cnn的最好效果达到了83.09 todo: redo -> lstm + attention todo: how to use relation embedding and ranking loss function todo: check the performance of pytorch implementation of multi-level cnn

10.22 加入pos信息,f1最好效果达到83.82 发现采用数据增强,实际运行效果大跌,猜测由于网络结构过于扁平,导致学习失败.

10.23 重新进行了预处理,统一设置word index, 便于更换word embedding 测试了lstm_with_attention的性能,实际最好效果74 测试双向lstm的基准性能,用于和lstm + attention进行比较, 实际效果为: 0.645左右 更换google词向量,发现最好效果为0.6418 ??猜测预处理有问题 simple_cnn 最好效果83.20 发现使用tensorflow, 无法得到可复现的结果 处理: 暂时不纠结预处理过程, 在google embedding的基础上继续处理 picewise-maxpooling 无效, 实际效果在250 * 4 cnn作用下,可以达到0.83, 不如直接采用maxpooling

11.27 bert 大法好,实际效果不过10,醉了~

12.6 rnf_cnn + elmo 84.5 todo: add pos_tag

semeval_2010_task8's People

Contributors

zhoudayang avatar

Stargazers

 avatar  avatar zhangxiaojia avatar  avatar  avatar xx_xavier avatar  avatar  avatar  avatar xyz avatar Lei avatar  avatar Liu Kai (刘 凯) / Leo avatar  avatar hoshi-hiyouga avatar 若谷 avatar f avatar

Watchers

 avatar

Forkers

wangvje

semeval_2010_task8's Issues

BERT效果不好,求解

我用bert作为sentence encoder之后效果也非常不好,准确率非常低,但是也有论文中使用bert做这个任务,且准确率很高,是我使用的方式不对吗?希望和您探讨一下这个问题。

preteat文件夹中找不到对应文件

周炀你好,我在执行你preteat和preteat1文件夹里面的代码时,有些文件找不到
例如:
google_words, google_lower_words, google_vec = read_vec("google_list.json", "google.npy")
glove_words, glove_lower_words, glove_vec = read_vec("glove_list.json", "glove.npy")
wiki_words, wiki_lower_words, wiki_vec = read_vec("wiki_fasttext.json", "wiki_fasttext.npy")
embedding = np.load("google.npy")
words = json.load(open("google_list.json"))
embedding = np.load("wiki.vec.npy")
words = json.load(open("words.json"))
这些google_list.json、google.npy等文件是需要我自己在网上下载吗?如果是需要我自己下载的话,可以把资源链接分享给我吗,我没找到对应的文件,不能正常的运行代码。十分感谢

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.