-
## Model structure of YOLOv8 detection models(P5) - `yolov8n/s/m/l/x`:
![YOLOv8_structure_v1.6](https://github.com/ultralytics/ultralytics/assets/27466624/57391d0f-1848-4388-9f30-88c2fb79233f)
C…
-
## Background
- `Should We Rely on Entity Mentions for Relation Extraction? Debiasing Relation Extraction with Counterfactual Analysis, NAACL2022` 논문에서 제시한 CoRE 기법이 RE Task 에서 높은 성능을 보였기 때문에 구현을 시도한…
-
Hi
I am using a GBT model with the quickscorer algorithm. For training, I create a yggdrasil_decision_forests::dataset::VerticalDataset and use the AppendExample interface for which I have to conve…
-
- 版本、环境信息
1)PaddleNLP和PaddlePaddle版本:PaddleNLP 2.4.0,PaddlePaddle2.3.1
2)系统环境:Windows,python3.8
- 复现信息:
```
python train.py \
--task_type relation_extraction \
--train_path student_da…
-
hi
I tried to use gallery-dl to download from (e621)
And i want to use the "parent_id" in (danbooru) i use
```
"parent-metadata": true,
```
and since (e621) and (danbooru) are really similar …
-
### 请提出你的问题
uie抽关系或者实体的时候能设置自动过滤么?比如probability低于0.5就不作为结果这样的?还是说全部抽出来后再另外处理?
我之前在百度的答疑群里看到有人回复说;
Taskflow里面通过probability已经过滤了一波,你可以设置自己的需求设置probability,例如你要求准确率高就设置大一点;
然后我就在加了这个参数,如下:
`en…
-
老师,为什么我用您的main训练出模型后,运行demo出错,AttributeError: 'NoneType' object has no attribute 'data'。
-
# 복수의 Model 조서 필요성
각각의 모델은 성능 한계가 존재하며 더욱 높은 성능을 위해 다른 모델을 시도할 필요가 있음
복수의 모델을 Ensemble 하여 성능을 높일 수 있음
# Baseline의 모델 정보
Baseline에서는 klue/bert-base를 사용 [Hugging Face](https://huggingface.co/klu…
-
### 请提出你的问题
Hello,首先必须要感谢 PaddlePaddle 团队提供的如此好用的 UIE。
我这边参照[这个教程](https://github.com/PaddlePaddle/PaddleNLP/blob/develop/applications/information_extraction/text/data_distill/README.md)进行微调后的 uie 模…
-
Hello, thanks for your excellent work.
I want to know how the candidate relations are produced. Are they from the origin dataset? If not, how you choose the candidates for each sample?