GithubHelp home page GithubHelp logo

Comments (8)

ymcui avatar ymcui commented on June 20, 2024

https://arxiv.org/abs/1806.00920
参考DRCD官方的论文,报BERT基线结果 F1 score 89.59% and Exact Match score 82.34%.

from chinese-bert-wwm.

zhanghan1992 avatar zhanghan1992 commented on June 20, 2024

能否提供一下阅读理解的实验代码,用bert官方代码单卡batch size开不到64

from chinese-bert-wwm.

ymcui avatar ymcui commented on June 20, 2024

参考:https://github.com/ymcui/CMRC2018-DRCD-BERT/
在这个基础上仅对lr_schedual进行了优化以提高基线效果。

from chinese-bert-wwm.

zhanghan1992 avatar zhanghan1992 commented on June 20, 2024

参考:https://github.com/ymcui/CMRC2018-DRCD-BERT/
在这个基础上仅对lr_schedual进行了优化以提高基线效果。

python run_cmrc2018_drcd_baseline.py
--vocab_file=${PATH_TO_BERT}/multi_cased_L-12_H-768_A-12/vocab.txt
--bert_config_file=${PATH_TO_BERT}/multi_cased_L-12_H-768_A-12/bert_config.json
--init_checkpoint=${PATH_TO_BERT}/multi_cased_L-12_H-768_A-12/bert_model.ckpt
--do_train=True
--train_file=${DATA_DIR}/cmrc2018_train.json
--do_predict=True
--predict_file=${DATA_DIR}/cmrc2018_dev.json
--train_batch_size=32
--num_train_epochs=2
--max_seq_length=512
--doc_stride=128
--learning_rate=3e-5
--save_checkpoints_steps=1000
--output_dir=${MODEL_DIR}
--do_lower_case=False
--use_tpu=False

论文效果是用这个配置跑的吗?

from chinese-bert-wwm.

ymcui avatar ymcui commented on June 20, 2024

请参考技术报告中的参数(batch 64, lr 3e-5,epoch 2,maxlen 512),未说明的代表使用的是默认配置。
另外,你贴的是multi-lingual BERT上的命令,不是Chinese BERT,do_lower_case需要设置为True,中文BERT是uncased模型。

from chinese-bert-wwm.

zhanghan1992 avatar zhanghan1992 commented on June 20, 2024

https://github.com/ymcui/CMRC2018-DRCD-BERT/ 的代码在32g v100上单卡batch size开不到64。论文里的实验在什么gpu上跑的?是通过多卡同步增大batch size吗?

from chinese-bert-wwm.

ymcui avatar ymcui commented on June 20, 2024

TPU v2

from chinese-bert-wwm.

ymcui avatar ymcui commented on June 20, 2024

reopen if necessary

from chinese-bert-wwm.

Related Issues (20)

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.