merge

This is a merge of pre-trained language models

Goal of this merge is to create good 8B roleplay model that will be cpable to RU, for people with low VRAM/RAM.

Tested on 800 replies, on various char cards, from very easy to hard ones.

This model is good on english RP/ERP, and acceptable on russian RP/ERP, it follow instructions, and enough creative.

While not properly tested, assistant is not bad too, both on ru and en.

What about template? Honestly, i don't know. it somehow works fine with LLama 3 instruct, L3 instruct names, L2 Chat, default from silly tavern, and even with ChatML. I liked modded alpaca and default.

Of course, this is 8B, so don't expect very much from it.

Downloads last month
6
Safetensors
Model size
8.03B params
Tensor type
FP16
·
Inference Providers NEW
This model is not currently available via any of the supported Inference Providers.

Model tree for OddTheGreat/Rraccoon_8B_v2

Collection including OddTheGreat/Rraccoon_8B_v2