30,000 AI models
too many really. But from what I read in conversations and posts I notice one thing: you all try out Model all the time and that’s fine, but I haven’t yet read that anyone habitually uses one Model over others. It seems like you use one template for a few days and then start with a new one. Don’t have your favorite? Which?
Introduce an interesting work: DARE (Drop And REscale)
DARE can merge multiple task-specific LLMs (e.g., WizardLM + WizardMath) into one LM with diverse abilities, ✅but without the need for retraining or GPUs
https://twitter.com/WizardLM_AI/status/1727672799391842468?t=alsj7WrhCzVzSxjN7vKOyQ&s=19