conflibertR: Inference and Fine-Tuning with 'ConfliBERT' Conflict Text Models

An interface to 'ConfliBERT', a pretrained language model for analyzing text about conflict and political violence (Hu et al. (2022) <doi:10.18653/v1/2022.naacl-main.400>). Provides functions for named entity recognition, binary and multilabel classification, and question answering, plus tools to fine-tune custom classifiers, compare several base model architectures, and run an interactive active-learning loop for efficiently labeling new data. Models are downloaded from 'Hugging Face' and run through the 'transformers' library for 'Python' via the 'reticulate' package.

Version: 0.5.3
Imports: cli, reticulate (≥ 1.34), tibble
Suggests: ggplot2, knitr, miniUI, rmarkdown, rstudioapi, shiny, testthat (≥ 3.0.0)
Published: 2026-07-10
DOI: 10.32614/CRAN.package.conflibertR (may not be active yet)
Author: Shreyas Meher ORCID iD [aut, cre]
Maintainer: Shreyas Meher <shreyasmeher at gmail.com>
BugReports: https://github.com/shreyasmeher/conflibertR/issues
License: MIT + file LICENSE
URL: https://github.com/shreyasmeher/conflibertR
NeedsCompilation: no
SystemRequirements: Python (>= 3.8); the Python modules 'torch', 'transformers' (>= 4.40), 'accelerate', 'peft', 'scikit-learn' and 'numpy', installable via conflibert_install()
Language: en
Citation: conflibertR citation info
Materials: README, NEWS
CRAN checks: conflibertR results

Documentation:

Reference manual: conflibertR.html , conflibertR.pdf
Vignettes: Active Learning with conflibertR (source, R code)
Getting Started with conflibertR (source, R code)

Downloads:

Package source: conflibertR_0.5.3.tar.gz
Windows binaries: r-devel: not available, r-release: not available, r-oldrel: not available
macOS binaries: r-release (arm64): conflibertR_0.5.3.tgz, r-oldrel (arm64): conflibertR_0.5.3.tgz, r-release (x86_64): not available, r-oldrel (x86_64): not available

Linking:

Please use the canonical form https://CRAN.R-project.org/package=conflibertR to link to this page.