Release of Codes - Biendata

Tsinghua University, Knowledge Engineering Group and Zhipu AI • $10,000 • 210 Team • 243 participants

PST-KDD-2024

2024-03-20 - Launch

2024-06-01 - Team Merger Deadline

2024-06-07 - Close

Home Competitions

Fanjin Zhang

OAG-Challenge Team

Release of Codes

posted in PST-KDD-2024

2024-06-11 07:49

To be eligible for the possible awards, winners who place in the **Top 15 on the TEST leaderboard** must open-source their solutions. The public GitHub repository must include all code necessary to reproduce the results. **Placeholder repositories will not be accepted.** The submission deadline is **23:59 June 12, 2024, AOE Time**. Please reply your (**Track, Team Name, Rank, Github URL**) in this thread (https://www.biendata.xyz/forum/view_post_category/1034649).
 
- The authors are responsible for addressing any inquiry about their code.
- Please add ``README`` and provide sufficient information, including:
  - **Instructions**: Detailed steps and exact commands required to reproduce the submitted result.
  - **Method Introduction**: A brief overview of your methods.
  - Good examples are [[WhoIsWho-IND Codes]](https://github.com/THUDM/whoiswho-top-solutions/tree/main/incorrect_assignment_detection) [[AQA Codes]](https://github.com/THUDM/OAG-AQA) [[PST Codes]](https://github.com/THUDM/paper-source-trace/tree/main).

# Note!!!
## Please provide your model checkpoints if your fine-tune pre-trained models.
## It is highly recommended to upload your model to huggingface, followed by Baidu NetDisk and Google Drive.
## Please provide the inference scripts on TEST data and your training scripts on the training data separately.

23 comments

NulltoZero reply Fanjin Zhang

2024-06-15 07:05

Reply

0

我们已经在github仓库中上传了我们以往的检查点，推理注意事项已在readme文件中说明，非常抱歉造成困扰。

NulltoZero reply Fanjin Zhang

2024-06-15 05:48

Reply

0

您好。为了保证我们审核代码的测试成功，由于transformers库的版本问题，我们提供的检查点是测试代码时一次性得出的。但在以往提交中，因为正负样本的随机采样，我们挑选的是本地训练Loss的最低的一次实验，以及根据提交得分调整不同策略的集成权重。非常抱歉带来困惑，但我们能够保证我们方案的真实性。

OAG-Challenge Team

Fanjin Zhang reply NulltoZero

2024-06-15 03:08

Reply

0

你好，我们运行你们提供的推理代码，测试集提交结果为 0.46129。有哪些情况可能导致和排行榜分数有差异呢？

glin reply bzsy

2024-06-14 11:14

Reply

0

对requirements.txt进行了修改并进行了本地测试，readme也备注了推理函数

glin reply bzsy

2024-06-14 07:14

Reply

0

好的好的，我今天之内处理

NulltoZero reply bzsy

2024-06-14 06:18

Reply

0

非常抱歉，我已经修改了文件路径。同时我在readme文件中添加了BAAI/bge-m3预训练模型下载位置。如果您想通过本地读取的方式加载bgem3模型，请修改对应文件读取位置。感谢！

bzsy reply NulltoZero

2024-06-14 04:17

Reply

0

您好，我是比赛的组织者之一。我们在运行您的代码bash test.sh时，在bert_spacy.py第550行，bert_lxe.py第1030行，bert_ab_dic.py第535行均出现OSError: Incorrect path_or_model_id: '/data/zsp/KDD_PST_2341/models/sci-roberta'. 您是否忘记修改绝对路径？在我们将路径修改至您在README文件中提到的/out/kddcup/sciroberta/文件夹后，此三个文件又出现新的报错：OSError: /XXXXX/KDDCUP2024_PST_Review-main/out/kddcup/sciroberta/ does not appear to have a file named config.json. 此外，您的模型是否需要下载BAAI/bge-m3预训练模型？如果需要，请提供一个放置预训练模型的文件夹位置以便手动下载并且放置模型文件。感谢。

Cristiano reply bzsy

2024-06-13 15:49

Reply

0

你好，我更新了requirements.txt，麻烦从github重新获取requirements.txt安装依赖。

Cristiano reply bzsy

2024-06-13 15:36

The comment has been deleted by author.

Jindou Wu reply bzsy

2024-06-13 11:52

Reply

0

已经添加requirements.txt 依赖，推理见infer_run.sh

edgeknife

2024-06-13 11:46

🥺data reply 🥺data

2024-06-13 11:34

Reply

0

I fixed some mistakes.

NulltoZero

2024-06-13 10:49

masatoh

2024-06-13 06:40

luxuantao reply bzsy

2024-06-13 09:53

Reply

0

更新下transformers版本，requirements.txt中已修正

Tilbur

2024-06-13 09:51

zxl_1999 reply bzsy

2024-06-13 09:42

Reply

0

Hello, now I have updated requirements.txt，had ensured that the configuration files only involve third-party modules required for the code to run.Please try to run it again, thx

bzsy reply zxl_1999

2024-06-13 09:12

Reply

0

您好，我是比赛的组织者之一。使用您提供的requirements.txt无法正确安装包。报错如下： Invalid requirement: 'aiohttp @ file:///home/conda/feedstock_root/build_artifacts/aiohttp_1701099469104/work' It looks like a path. Does it exist ?

bzsy reply luxuantao

2024-06-13 08:55

Reply

0

您好，我按照您提供的README文件配置了环境并且放置模型文件后，运行python inference.py出现报错： RuntimeError: Error(s) in loading state_dict for BertForSequenceClassification: Missing key(s) in state_dict: "bert.embeddings.position_ids".

bzsy reply luxuantao

2024-06-13 08:55

Reply

0

您好，我按照您提供的README文件配置了环境并且放置模型文件后，运行python inference.py出现报错： RuntimeError: Error(s) in loading state_dict for BertForSequenceClassification: Missing key(s) in state_dict: "bert.embeddings.position_ids".

bzsy reply Jindou Wu

2024-06-13 08:30

Reply

0

您好，您的代码中并没有包含requirements.txt或有关所需要用到的包的任何文件，因此无法运行您的代码。请补充一份相关文件以便对您的结果进行检查。另外，请提供可以直接进行预测（而非训练后再进行预测）的代码或命令行文件。

bzsy reply Jindou Wu

2024-06-13 08:30

The comment has been deleted by author.

Jindou Wu reply bzsy

2024-06-13 08:13

Reply

0

已经通过，麻烦再查看下

bzsy reply Cristiano

2024-06-13 07:58

Reply

0

您好，我是比赛的组织者之一。您的requirements.txt中存在包冲突的现象，无法正确配置环境。报错如下： ERROR: Cannot install -r requirements.txt (line 2), -r requirements.txt (line 200), -r requirements.txt (line 220), -r requirements.txt (line 292), -r requirements.txt (line 333), -r requirements.txt (line 340), -r requirements.txt (line 41), -r requirements.txt (line 57), -r requirements.txt (line 65) and torch==2.2.0 because these package versions have conflicting dependencies. The conflict is caused by: The user requested torch==2.2.0 accelerate 0.27.2 depends on torch>=1.10.0 cogdl 0.6 depends on torch deepspeed 0.14.2 depends on torch effdet 0.4.1 depends on torch>=1.12.1 ogb 1.3.6 depends on torch>=1.6.0 peft 0.11.1 depends on torch>=1.13.0 sentence-transformers 2.2.2 depends on torch>=1.6.0 timm 0.9.16 depends on torch torchaudio 2.1.2 depends on torch==2.1.2 To fix this you could try to: 1. loosen the range of package versions you've specified 2. remove package versions to allow pip attempt to solve the dependency conflict ERROR: ResolutionImpossible: for help visit https://pip.pypa.io/en/latest/topics/dependency-resolution/#dealing-with-dependency-conflicts

Jindou Wu reply Jindou Wu

2024-06-13 07:22

The comment has been deleted by author.

🥺data

2024-06-13 06:51

bzsy reply glin

2024-06-13 06:32

Reply

0

您好，我是比赛的组织者之一。您提供的requirements.txt文件中有着大量的python包，我们在验证过程中有很多包存在着版本问题无法正确下载。请您提供正确的并且仅包含必要包的requirements.txt文件。感谢。

bzsy reply Jindou Wu

2024-06-13 06:29

Reply

1

您好，我是比赛的组织者之一。您上传在谷歌云盘的模型需要权限才能访问，我已经发送了申请，请通过，感谢。

luxuantao reply bzsy

2024-06-13 06:27

Reply

0

已修正，thx

bzsy reply luxuantao

2024-06-13 02:53

Reply

0

您好，我是比赛的组织者之一。运行您的代码python inference.py会出现报错：huggingface_hub.utils._validators.HFValidationError: Repo id must be in the form 'repo_name' or 'namespace/repo_name': '/mnt/nlp-ali/usr/tiance/PLM/scibert_scivocab_uncased'. Use `repo_type` argument if needed. 您是否忘记修改模型路径？另外请在README中标注在哪个位置放置您模型的checkpoint。感谢。

bzsy reply anaivebird

2024-06-13 02:41

Reply

0

您好，我是比赛的组织者之一。在我运行run_test.py后会报错：FileNotFoundError: [Errno 2] No such file or directory: 'XXX/kdd_pst/data/paper_source_trace_2023_final_filtered.json' 这个json文件我们似乎并没有提供，在您提到的数据集中也并没有找到。

cklwanfifa

2024-06-12 16:41

AoboSama

2024-06-12 15:39

AoboSama

2024-06-12 15:26

The comment has been deleted by author.

heart11

2024-06-12 14:02

zxl_1999

2024-06-12 13:12

gorgeousdays

2024-06-12 11:38

Jindou Wu

2024-06-12 09:15

Cristiano

2024-06-12 07:30

anaivebird reply Fanjin Zhang

2024-06-12 07:08

Reply

0

already fixed, thanks

luxuantao

2024-06-12 07:04

OAG-Challenge Team

Fanjin Zhang reply anaivebird

2024-06-12 03:25

Reply

0

Hi, Please provide your model checkpoints. Please provide the inference scripts on TEST data and your training scripts on the training data separately. The current codebase seems work on validation set.

anaivebird

2024-06-12 01:03

glin

2024-06-11 16:34