This is Wizard 8x22, quantized to GPTQ with these parameters:

python3 quant.py alpindale/WizardLM-2-8x22B /workspace/wizard-4bit custom --bits 4 --group_size 128 --desc_act 1 --damp 0.1 --dtype float16 --seqlen 16384 --num_samples 256 --cache_examples 0 --trust_remote_code

The dataset used was openerotica/erotiquant2. I have included a script to reconstitute the files into one.