3.2bpw/h6 exl2 quantization of
Undi95/Miqu-MS-70B using default exllamav2 calibration dataset.
This is a merge of pre-trained language models created using
mergekit.
Feedback on this model is greatly appreciated! I hope this new merge method will be able to fill some hole Miqu have.
Since it was made with model using different prompt format, the following should work.
### Instruction:
{system prompt}
### Input:
{prompt}
### Response:
{output}
SYSTEM: <ANY SYSTEM CONTEXT>
USER:
ASSISTANT:
This model was merged using the
Model Stock merge method using
152334H/miqu-1-70b-sf as a base.
1models:
2 - model: NeverSleep/MiquMaid-v2-70B
3 - model: sophosympatheia/Midnight-Miqu-70B-v1.0
4 - model: migtissera/Tess-70B-v1.6
5 - model: 152334H/miqu-1-70b-sf
6merge_method: model_stock
7base_model: 152334H/miqu-1-70b-sf
8dtype: bfloat16
If you want to support me, you can
here.
