Airoboros 33b GPT4 1.2 merged with kaiokendev's
33b SuperHOT 8k LoRA, quantised using GPTQ-for-LLaMa.
To easily use this model, you can use Oobabooga's
Text Generation WebUI and run it with the
--monkeypatch flag (and use the Exllama loader for best speeds. Note this must be manually installed unless you use the 1 click installer.)
