qiufengyuyi / sequence_tagging Goto Github PK
View Code? Open in Web Editor NEWusing bilstm-crf,bert and other methods to do sequence tagging task
using bilstm-crf,bert and other methods to do sequence tagging task
运行run_torch的时候数据处理出现问题,一是data_processing中没有lstmcrf_prepare_data这个文件,另外训练语料通过不同分词器得到的word segment的json文件结构有问题,是不完整的,需要把每个json文件的最后一条不完整的数据删除掉
in train_helper.py
from optimization import BertAdam
i can't find BertAdam in optimization
only class AdamWeightDecayOptimizer in optimization.py
评估或者预测模式的时候 get_rnn_cell函数中dropout_rate参数一直是0.75
Two problems:
dice_dsc_loss
or focal_dsc_loss
which portion i have to amend?
https://github.com/qiufengyuyi/sequence_tagging/blob/master/models/bert_mrc.py
然后看到代码中关于bert_mrc计算f1时,不应该是根据实体来计算的吗
看代码中是根据start end指针来分开计算的吗
最近也在复现bert-mrc模型,基于MSRA数据集,但是效果10个epoch F1只能达到 0.64,50个epoch效果会好点。但是也到不了0.9多,不知道是哪里存在问题了。
然后在根据start end指针提取实体的时候,会遇到start指针有些不能与end指针匹配上的问题。
In train.py
, the function run_lan
is not defined.
你好,我使用新数据训练 发现train dev的loss 很小 只有 1e-14 这个级别 这个应该不正常吧 是哪里出问题了呢
请问作者是不是没有把train_helper.py
里面的if args.do_test:
部分写进去呢?运行load_and_predict.py
时,显示No such file or directory:'prediction_result_completelabel_bert_nocrf_dlloss.npy'
,应该是没有进行预测过程生成.npy
文件吧
嗨,在试运行BertForMWA时出现文件缺失的情况
具体报错:No module named 'models.layers.crf_layers_pytorch'
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google ❤️ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.