Links to models.

[Ph0rk0z]

Opt (oldGPTQ):
https://huggingface.co/autobots/opt-6b-4-bit

Pygmalion (oldGPTQ)
https://huggingface.co/OccamRazor/pygmalion-6b-dev-gptq-4bit
https://huggingface.co/OccamRazor/pygmalion-6b-gptq-4bit

OSST (oldGPTQ)
Tested and working.
https://drive.google.com/file/d/1lAlID8RwFKOEvQSHt6owcg9hGvBSc5EY/view?usp=share_link

NeoXT-Chatbot-Base (oldGPTQ)
Did not work in ooba.. will try in kobold or if I get autograd working.
https://drive.google.com/file/d/1OaKKqUFVffExi5fGezwYrMjG6Ac3romV/view?usp=share_link

Llama (oldGPTQ)
https://huggingface.co/decapoda-research/llama-13b-hf-int4/tree/main
https://huggingface.co/decapoda-research/llama-7b-hf-int4/tree/main
https://huggingface.co/maderix/llama-65b-4bit/tree/main

[Ph0rk0z]

Lotus 12b in GPTQv2 with act order + true sequential
https://huggingface.co/autobots/lotus-12b-gptqv2-4bit/tree/main

I had to stop converting the V1 model and then the next time it wouldn't convert anymore. Will have to try again.

Pythia 12b GPTQv2

https://huggingface.co/autobots/pythia-12b-gptqv2-4bit

Gpt-4chan in 4bit GPTQv2
https://huggingface.co/autobots/gpt-j-fourchannel-4bit