After 500+ LoRAs made, here is the secret

RelatedInsightsHighlights

The difference between "pretty good" and "great data" is hundreds of millions of dollars in wasted parameters. Basically every "great" model today, can be 10x'd with better data curation. How much this matters is genuinely shocking, and everyone is ignoring it. https://t.co/kJrZA56mQu

Will Manidis

x.com

Here is a great machine learning exercise: 1. Pick a dataset (any dataset) and build a simple model. 2. Once you have it working, don't touch the model again. From that point on, focus on improving the data: 1. Fix any label issues 2.... See more

Santiago x.com

Thumbnail of www-x-com-abacaj-status-1709798110153646084-6081a12305a84aa8

I couldn’t agree more, this is the secret https://t.co/hnbjfvLzQ9

anton

x.com