Ayan

> > > Anyone found a solution? I couldn't find one. I encountered this problem as well, resolved by reinstalling 64bit version of python.

If I understood correctly, you will need to compute style embedding arrays using style encoder and use Principal Component Analysis on that data to convert each style embedding into a...

From my experience of changing to llama-2 you'll need to rewrite all the prompt templates for it to work more or less stably. The code is still VERY brittle.

I have a [fork](https://github.com/Miru302/generative_agents) that sometimes runs. It doesn't work well, often errors out because output parsing often breaks. Innitially, used another PR(#64), not this one. You can clone that,...