Practical overview
When it fits
Use BOND when you need repository-provided workflows and datasets for experimenting with open-domain NER under distant or weak supervision, especially the documented CoNLL03 self-training or baseline paths.
Who it is for
- Researchers evaluating open-domain named-entity recognition methods.
- Practitioners prototyping NER training with weakly or distantly labeled datasets.
Documented uses
- Train a NER model with the BOND self-training workflow.
- Train and evaluate the Stage I baseline on CoNLL03.
- Evaluate NER results across the five included datasets.
Requirements & limits
The repository is public and identifies its license as Apache-2.0.
- The stated software environment is Python 3.7, PyTorch 1.3 and Hugging Face Transformers 2.3.0.
- Five preprocessed open-domain distantly or weakly labeled NER datasets are included.
- Gazetteer information and distant-label generation code are available by contacting the email listed in the repository, rather than through the included dataset files.
Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.
Run the documented CoNLL03 training and evaluation workflow
Use the repository's BOND self-training workflow for CoNLL03.
Before you start
- A Linux environment.
- Python 3.7, PyTorch 1.3 and Hugging Face Transformers 2.3.0.
- The BOND repository.
Steps
Change to the BOND repository directory.
Run
./scripts/conll_self_training.sh.
Expected outcome
The documented workflow trains and evaluates BOND on CoNLL03.
Guide limitations
- The repository documents this command for CoNLL03; it does not establish compatibility with other software versions.
Run the CoNLL03 Stage I baseline
Train and evaluate the repository's Stage I baseline on CoNLL03.
Before you start
- A Linux environment.
- Python 3.7, PyTorch 1.3 and Hugging Face Transformers 2.3.0.
- The BOND repository.
Steps
Change to the BOND repository directory.
Run
./scripts/conll_baseline.sh.
Expected outcome
The documented Stage I baseline training and evaluation workflow runs on CoNLL03.
Guide limitations
- The repository documents this command for CoNLL03; it does not establish compatibility with other software versions.
Sources & review
Source record details
- Source availability
- Active
- Sources checked
- 17 October 2026
- Original website title
- cliang1453/BOND · Official homepage title · retained in its source language