dragonfarm: Fine-Tune Small Language Models with LoRA from R

Fine-tune small (100M to 3B parameter) causal language models with LoRA (Low-Rank Adaptation) from R. Datasets are mapped to chat-format prompts and responses, training runs in a background 'Python' process built on Hugging Face 'transformers' and 'peft', and a 'shiny' app offers drag-and-drop dataset upload and column mapping. 'Python' dependencies are declared through 'reticulate' and resolved automatically on first use.

Version:	0.3.3
Depends:	R (≥ 4.1)
Imports:	bslib (≥ 0.6.0), cli, glue, jsonlite, plotly, processx, ps, reticulate (≥ 1.41.0), rlang, shiny (≥ 1.8.0), sortable, stats, tools, utils, withr, zip
Suggests:	arrow, ellmer, httr2, knitr, pkgload, rmarkdown, testthat (≥ 3.0.0)
Published:	2026-10-07
DOI:	10.32614/CRAN.package.dragonfarm (may not be active yet)
Author:	Tejas Patel [aut, cre]
Maintainer:	Tejas Patel <algocrat at gmail.com>
BugReports:	https://github.com/tejas4patel/dragon-farm/issues
License:	MIT + file LICENSE
URL:	https://github.com/tejas4patel/dragon-farm
NeedsCompilation:	no
SystemRequirements:	'Python' (>= 3.10). The 'uv' tool is installed automatically by 'reticulate' to build the 'Python' environment.
Materials:	README, NEWS
CRAN checks:	dragonfarm results

Documentation:

Reference manual:	dragonfarm.html , dragonfarm.pdf
Vignettes:	The post-training loop: fine-tune, judge, prefer, repeat (source, R code) Quickstart: fine-tune a small model from R (source, R code) Reinforcement learning with rewards you can check (source, R code) The dragon-farm app (source, R code)

Downloads:

Package source:	dragonfarm_0.3.3.tar.gz
Windows binaries:	r-devel: not available, r-release: not available, r-oldrel: not available
macOS binaries:	r-release (arm64): not available, r-oldrel (arm64): not available, r-release (x86_64): not available, r-oldrel (x86_64): not available

Linking:

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