http://simonwillison.net/2023-03-17T23:08:40+00:00Simon Willison2023-03-17T23:08:40+00:002023-03-17T23:08:40+00:00https://simonwillison.net/2023/Mar/17/fine-tune-llama-to-speak-like-homer-simpson/#atom-tag

<p><strong><a href="https://replicate.com/blog/fine-tune-llama-to-speak-like-homer-simpson">Fine-tune LLaMA to speak like Homer Simpson</a></strong></p> Replicate spent 90 minutes fine-tuning LLaMA on 60,000 lines of dialog from the first 12 seasons of the Simpsons, and now it can do a good job of producing invented dialog from any of the characters from the series. This is a really interesting result: I’ve been skeptical about how much value can be had from fine-tuning large models on just a tiny amount of new data, assuming that the new data would be statistically irrelevant compared to the existing model. Clearly my mental model around this was incorrect. <p>Tags: <a href="https://simonwillison.net/tags/the-simpsons">the-simpsons</a>, <a href="https://simonwillison.net/tags/ai">ai</a>, <a href="https://simonwillison.net/tags/generative-ai">generative-ai</a>, <a href="https://simonwillison.net/tags/llama">llama</a>, <a href="https://simonwillison.net/tags/local-llms">local-llms</a>, <a href="https://simonwillison.net/tags/llms">llms</a>, <a href="https://simonwillison.net/tags/replicate">replicate</a>, <a href="https://simonwillison.net/tags/fine-tuning">fine-tuning</a></p>

2008-11-19T16:48:24+00:002008-11-19T16:48:24+00:00https://simonwillison.net/2008/Nov/19/pownce/#atom-tag

<p><strong><a href="http://pownce.com/leahculver/notes/4231180/">Lisa Simpson - crossword fan and ... Django developer?</a></strong></p> The Django Pony strikes again. <p>Tags: <a href="https://simonwillison.net/tags/django">django</a>, <a href="https://simonwillison.net/tags/djangopony">djangopony</a>, <a href="https://simonwillison.net/tags/leah-culver">leah-culver</a>, <a href="https://simonwillison.net/tags/lisasimpson">lisasimpson</a>, <a href="https://simonwillison.net/tags/the-simpsons">the-simpsons</a></p>