Comments (2)
您好,WN18RR的hits@10表现也比较不稳定, UMLS和FB15k-237的Hits@10比较稳定。可参考 #9 。
from kg-bert.
@yao8839836 你好,我运行run_bert_relation_prediction.py得到的结果与论文结果相去甚远,是参数设置的不对吗
02/04/2021 18:31:46 - INFO - main - ***** Test results *****
02/04/2021 18:31:46 - INFO - main - acc = 0.6901355995327657
02/04/2021 18:31:46 - INFO - main - eval_loss = 0.7939625151198486
02/04/2021 18:31:46 - INFO - main - global_step = 301980
02/04/2021 18:31:46 - INFO - main - loss = 0.3695593820695463
我的参数输入
--task_name kg --do_train --do_eval --do_predict --data_dir ./data/FB15K --bert_model bert-base-cased --max_seq_length 25 --train_batch_size 32 --learning_rate 5e-5 --num_train_epochs 20.0 --output_dir ./output_FB15K/ --gradient_accumulation_steps 1 --eval_batch_size 512
原代码读取entity2text.txt
文件报字符错误,我换成了'utf'格式
with open(os.path.join(data_dir, "entity2text.txt"), 'r', encoding='utf') as f:
ent_lines = f.readlines()
for line in ent_lines:
temp = line.strip().split('\t')
ent2text[temp[0]] = temp[1]
if data_dir.find("FB15") != -1:
with open(os.path.join(data_dir, "entity2text.txt"), 'r', encoding='utf') as f:
ent_lines = f.readlines()
for line in ent_lines:
temp = line.strip().split('\t')
#first_sent_end_position = temp[1].find(".")
ent2text[temp[0]] = temp[1]#[:first_sent_end_position + 1]
另外由于网络环境问题,在线下载模型比较慢,我手动换成读取已经下载好的本地模型bert-base-cased-vocab.txt
和bert-base-cased.tar.gz
tokenizer = BertTokenizer.from_pretrained("D:\\Documents\\code\\pretrained\\bert\\bert-base-cased-vocab.txt", do_lower_case=args.do_lower_case)
model = BertForSequenceClassification.from_pretrained("D:\\Documents\\code\\pretrained\\bert\\bert-base-cased.tar.gz",
cache_dir=cache_dir,
num_labels=num_labels)
希望能得到你的帮助
from kg-bert.
Related Issues (20)
- commands not working for prediction HOT 1
- epoch not working, pytorch issue? HOT 1
- list index out of error HOT 3
- estimated prediction time? HOT 1
- 是否支持中文 HOT 2
- GPU的使用 HOT 2
- Prediction Score for each time is different
- 论文中模型图
- hit@10在LP任务的testing阶段一直为0
- Using this with roberta-base
- gpu的使用
- UnicodeDecodeError: 'charmap' codec can't decode byte 0x90 in position 1519: character maps to <undefined>
- Time complexity of test triple ranking HOT 2
- Pre trained model
- 请问该文发表到正式期刊或会议了吗? HOT 2
- why umls dataset does not use long text? HOT 1
- 关于FB15K-237数据集文本句子的问题 HOT 2
- Run the model over a different kg
- 作者你好 请问下您的这篇论文现在发布在哪篇期刊上了? HOT 2
- 作者您好,我在您的实验的基础上将BERT模型替换为了DistilBERT模型,运行fb15k-237数据集,但是ACC为0,您知道是什么原因嘛
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
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.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from kg-bert.