Spaces:
Build error
A newer version of the Gradio SDK is available:
5.23.2
title: Augmented Poetry
emoji: ⚡
colorFrom: red
colorTo: gray
sdk: gradio
sdk_version: 3.4
app_file: app.py
pinned: false
- fine-tune a large language model (LLM) using the text corpus of a specific poet
select a certain rhyme from the gutenberg corpus and fine-tune on this
try fine-tuning on a few lines of a poem that Eva has started
run in a docker container and transfer to another machine
Is it better to have a sequence to sequence transformer trained on sucessive lines of the poetry corpus??
https://github.com/huggingface/transformers/tree/main/examples/pytorch/language-modeling
merve/poetry only has 573 rows.
TODO: - upload the gutenberg poetry corpus up to huggingface - ask the lady who made it
Research
https://github.com/aparrish/gutenberg-dammit/ implement language generation with a basic transformer
https://github.com/aparrish/gutenberg-poetry-corpus Gutenberg Poetry Autocomplete, a search engine-like interface for writing poems mined from Project Gutenberg. (A poem written using this interface was recently published in the Indianapolis Review!)
https://ymeadows.com/en-articles/fine-tuning-transformer-based-language-models https://thegradient.pub/prompting/ https://towardsdatascience.com/fine-tuning-for-domain-adaptation-in-nlp-c47def356fd6 https://ruder.io/recent-advances-lm-fine-tuning/