Duplicate from nbeerbower/gutenberg2-dpo

Co-authored-by: Nicholas Beerbower <[email protected]>

.gitattributes

@@ -0,0 +1,58 @@
+*.7z filter=lfs diff=lfs merge=lfs -text
+*.arrow filter=lfs diff=lfs merge=lfs -text
+*.bin filter=lfs diff=lfs merge=lfs -text
+*.bz2 filter=lfs diff=lfs merge=lfs -text
+*.ckpt filter=lfs diff=lfs merge=lfs -text
+*.ftz filter=lfs diff=lfs merge=lfs -text
+*.gz filter=lfs diff=lfs merge=lfs -text
+*.h5 filter=lfs diff=lfs merge=lfs -text
+*.joblib filter=lfs diff=lfs merge=lfs -text
+*.lfs.* filter=lfs diff=lfs merge=lfs -text
+*.lz4 filter=lfs diff=lfs merge=lfs -text
+*.mlmodel filter=lfs diff=lfs merge=lfs -text
+*.model filter=lfs diff=lfs merge=lfs -text
+*.msgpack filter=lfs diff=lfs merge=lfs -text
+*.npy filter=lfs diff=lfs merge=lfs -text
+*.npz filter=lfs diff=lfs merge=lfs -text
+*.onnx filter=lfs diff=lfs merge=lfs -text
+*.ot filter=lfs diff=lfs merge=lfs -text
+*.parquet filter=lfs diff=lfs merge=lfs -text
+*.pb filter=lfs diff=lfs merge=lfs -text
+*.pickle filter=lfs diff=lfs merge=lfs -text
+*.pkl filter=lfs diff=lfs merge=lfs -text
+*.pt filter=lfs diff=lfs merge=lfs -text
+*.pth filter=lfs diff=lfs merge=lfs -text
+*.rar filter=lfs diff=lfs merge=lfs -text
+*.safetensors filter=lfs diff=lfs merge=lfs -text
+saved_model/**/* filter=lfs diff=lfs merge=lfs -text
+*.tar.* filter=lfs diff=lfs merge=lfs -text
+*.tar filter=lfs diff=lfs merge=lfs -text
+*.tflite filter=lfs diff=lfs merge=lfs -text
+*.tgz filter=lfs diff=lfs merge=lfs -text
+*.wasm filter=lfs diff=lfs merge=lfs -text
+*.xz filter=lfs diff=lfs merge=lfs -text
+*.zip filter=lfs diff=lfs merge=lfs -text
+*.zst filter=lfs diff=lfs merge=lfs -text
+*tfevents* filter=lfs diff=lfs merge=lfs -text
+# Audio files - uncompressed
+*.pcm filter=lfs diff=lfs merge=lfs -text
+*.sam filter=lfs diff=lfs merge=lfs -text
+*.raw filter=lfs diff=lfs merge=lfs -text
+# Audio files - compressed
+*.aac filter=lfs diff=lfs merge=lfs -text
+*.flac filter=lfs diff=lfs merge=lfs -text
+*.mp3 filter=lfs diff=lfs merge=lfs -text
+*.ogg filter=lfs diff=lfs merge=lfs -text
+*.wav filter=lfs diff=lfs merge=lfs -text
+# Image files - uncompressed
+*.bmp filter=lfs diff=lfs merge=lfs -text
+*.gif filter=lfs diff=lfs merge=lfs -text
+*.png filter=lfs diff=lfs merge=lfs -text
+*.tiff filter=lfs diff=lfs merge=lfs -text
+# Image files - compressed
+*.jpg filter=lfs diff=lfs merge=lfs -text
+*.jpeg filter=lfs diff=lfs merge=lfs -text
+*.webp filter=lfs diff=lfs merge=lfs -text
+# Video files - compressed
+*.mp4 filter=lfs diff=lfs merge=lfs -text
+*.webm filter=lfs diff=lfs merge=lfs -text

README.md

@@ -0,0 +1,64 @@
+---
+language:
+- en
+license: apache-2.0
+---
+![image/png](https://huggingface.co/nbeerbower/gemma2-gutenberg-27B/resolve/main/gutenberg2-header.png?download=true)
+
+# Gutenberg2 DPO
+
+A DPO dataset meant to enhance the writing capabilities of LLMs using public domain books from [Project Gutenberg](https://gutenberg.org/).
+
+Inspired by Jon Durbin's Gutenberg DPO dataset: [jondurbin/gutenberg-dpo-v0.1](https://huggingface.co/datasets/jondurbin/gutenberg-dpo-v0.1)
+
+## Process
+
+Various books were selected from Project Gutenberg by personal preference and recommendation by Claude 3.5 Sonnet.
+
+They were then parsed by chapter using [chapterize](https://github.com/JonathanReeve/chapterize). Books that failed to parse or had mangled results were dropped from consideration.
+
+Chapters were then manually edited to remove: illustration tags, notes, etc. Superfluous line breaks were removed with a regular expression. 
+
+Each chapter had a summary generated by an LLM. The summary of the previous chapter and the text from the current chapter were then used to generate a prompt that would result in an LLM writing that chapter.
+
+The prompts were then used to generate a "rejected" value, while the original chapter text was used as "chosen."
+
+[nbeerbower/mistral-nemo-bophades-12B](https://huggingface.co/nbeerbower/mistral-nemo-bophades-12B) was used to generate summaries, prompts, and rejected values.
+
+## Code
+
+[gutenberg-tools](https://github.com/nbeerbower/gutenberg-tools) is MIT licensed on GitHub.
+
+## Acknowledgements
+
+- __Jon Durbin__ for the original idea and process, thank you for your contributions to the open source community
+- __Jonathan Reeve__ for the chapterize library
+- __Mistral AI__ for their state of the art, permissively licensed, open LLMs
+- __Project Gutenberg__ and all their volunteers, thank you for preserving literary history and making it available to all
+- __koboldcpp__ for providing a fast and easy way to run LLM inference (shoutout to the KoboldAI Discord)
+- __llama.cpp__ and Georgi Gerganov for GGUF
+- __Michael Radermacher__ for quantizing my models and especially the one used to generate this dataset
+
+Dedicated to my late grandfather, George F. Beerbower, an engineer and avid reader.
+
+## Selected Works
+
+- Alice's Adventures in Wonderland - Lewis Carroll 
+- Around the World in Eighty Days - Jules Verne
+- The Awakening - Kate Chopin
+- The Canterville Ghost - Oscar Wilde
+- The Invisible Man - H.G. Wells
+- The Island of Dr. Moreau - H.G. Wells
+- The Metamorphosis - Franz Kafka
+- Peter Pan - J.M. Barrie
+- The Picture of Dorian Gray - Oscar Wilde
+- The Red Badge of Courage - Stephen Crane
+- The Secret Agent - Joseph Conrad
+- The Adventures of Tom Sawyer - Mark Twain
+- The Wonderful Wizard of Oz - L. Frank Baum
+
+## Disclaimer
+
+This dataset is provided for free and as-is with absolutely no warranty. Some data may contain errors and be periodically modified.
+
+You are free to modify and redistribute it as you wish.