Sao10K/Ana-v1-m7

Ana - A model solely focused on the RP / ERP Experience.

- Little to no Censorship during Roleplay. May be kinda censored at 0 Conext. Does not affect Roleplays.
- Little to no Positivity Bias during Roleplay, from my tests atleast.
- Verbose
- Kinda Smart
- Slightly Horny by default.

Purely experimental ----- I do not know what I am doing, this is for fun. A merge + train. Final qLoRA train took 3 hours on a 4090.

*Kinda forgot, but alpaca format works fine lol

Why no 70Bs? Because the 4090 I have currently broke down. falling back to 3060 12gb.. I'm runpod training them for now.