Practical overview
When it fits
Use Golgotha when you want to obtain transformer-based embeddings or tokenization from an R workflow and use the results in downstream NLP modelling.
Who it is for
- R developers building NLP workflows with transformer-based text representations.
- Practitioners preparing text for classification, regression, sentiment labelling, tagging, lemmatization, or similarity analysis.
Documented uses
- Generate sentence embeddings from text for downstream modelling.
- Create token-level embeddings or tokenize text for analysis.
- Prepare multilingual text representations for similarity or other NLP tasks.
Requirements & limits
- It is an R library that wraps Python Transformers through reticulate, rather than a hosted analysis interface.
- Transformers and PyTorch are required; reticulate is described as normally handling their setup.
- BERT examples omit an architecture argument, while other model architectures require one in the documented examples.
- The model list is not exhaustive and depends on the Transformers version.
- The public repository lists the MPL-2.0 license.
Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.
Install Golgotha and prepare a BERT model
Install the R package, download a BERT model, and load it for embedding or tokenization calls.
Before you start
- R
- The Python Transformers and PyTorch modules, connected through reticulate
Steps
In R, install the development version with devtools::install_github("bnosac/golgotha", INSTALL_opts = "--no-multiarch").
Load the package with library(golgotha).
Download a model, for example transformer_download_model("bert-base-multilingual-uncased").
Load it with transformer("bert-base-multilingual-uncased"); the README shows predict() calls with types embed-sentence, embed-token, and tokenise.
Expected outcome
A downloaded and loaded BERT model is ready for the documented sentence-embedding, token-embedding, or tokenization calls.
Guide limitations
- The guide describes the documented setup and example calls, not a tested execution.
- The package requires Transformers and PyTorch; available models depend on the installed Transformers version.
- For architectures other than BERT, the README says to supply an architecture argument.
Sources & review
Source record details
- Source availability
- Active
- Sources checked
- 18 October 2026
- Original website title
- bnosac/golgotha · Official homepage title · retained in its source language