Skip to content

Latest commit

 

History

History
85 lines (60 loc) · 2.61 KB

README.md

File metadata and controls

85 lines (60 loc) · 2.61 KB
Raw

NLCC 2024 Shared Task 2: Nominal Compound Chain Extraction

切换到中文

Welcome to the Nominal Compound Chain Extraction project for the NLCC 2024 Shared Task 2 competition. This project contains the necessary files and instructions to train, validate, and test a model for extracting nominal compound chains from given datasets.

Project Structure

.
├── data
│   └── nlpcc_data
│       ├── train.json
│       ├── valid.json
│       └── test.json
│   └── submit
│   └── save
├── src
├── inference.py
├── main.py
├── README.md
└── requirements.txt
  • data/: Directory containing the dataset.
    • nlpcc_data/: Subdirectory containing the train, validation, and test JSON files.
    • save/: Subdirectory containing the model weights.
    • submit/: Subdirectory containing the predicted JSON files.
  • src: Directory containing the source code.
  • inference.py: Script to load the best performing model and obtain prediction results on the test set.
  • main.py: Script to train the model and obtain the best model on the validation set.
  • README.md: This file.
  • requirements.txt: File containing the required Python packages.

Dataset

  • train.json: Contains the training data with corpus and labels.
  • valid.json: Contains the validation data with corpus and labels.
  • test.json: Contains the test data with document ID and corpus, but without labels.

Setup

  1. Clone the repository to your local machine.

  2. Navigate to the project directory.

  3. Install the required packages using the following command:

    pip install -r requirements.txt

Training the Model

To train the model and obtain the best model on the validation set, run the following command:

python main.py

The best model will be saved in the data/save/ directory.

Inference

To load the best performing model and obtain the prediction results on the test set, run the following command:

python inference.py --chk best_12.pth.tar

Please replace best_12.pth.tar with the name of the best model file.

The prediction results will be saved in the data/submit/submit.jsonl file.

Submission

After obtaining the prediction results, you can zip the submit.jsonl file and upload the zippped file to the competition platform to obtain your score.

License

This project is licensed under the Apache License 2.0 License. See the LICENSE file for more details.

Contact

For any issues or questions, please contact us.

Good luck with the competition!