AlekseyCalvin/Poetry_Categorized_Chat_via_schifferlearning

---
license: apache-2.0
---
## POETRY CATEGORIZED (LLM Chat Edition)

This is my chat templatted adaptation of the [Poetry Categorized](https://huggingface.co/datasets/schifferlearning/Poetry-Categorized) dataset by [Theo Schifferli](https://github.com/theoschiff) (aka [schifferlearning](https://huggingface.co/schifferlearning)). <br>
It is a moderately scoped anthology of poems from the English-language literary canon, organized by author name, title, verse-form (very broadly classified between just four categories), and stanza-sequence information (where a poem starts and ends). <br>
The original dataset is just about metadata-free, as far as I can tell. As such, I can only speculate as to the original source of this corpus. <br>
But the composition/scope of the material suggests it was probably sourced from Project Gutenberg, either directly, or maybe via one of Allison Parrish's [alchemies thereof](https://github.com/aparrish/gutenberg-poetry-corpus). <br>
This version is not quite lossless apropos the original [Poetry Categorized](https://huggingface.co/datasets/schifferlearning/Poetry-Categorized), as it abridges poem titles and the stanza sequence information. <br>
Granted, I do add a hefty system prompt with a much more fine-grained breakdown of the original dataset's three formal verse categories (Quatrain, Sonnet, and Octave). <br>
With that said, somehow integrating the structural sequencing info may have benefitted this adaptation... So, I might put together another version soon. <br>