mini-magnum-12b-v1.1 - EXL2 8bpw max

This is a 8bpw EXL2 quant of intervitens/mini-magnum-12b-v1.1

This quant was made using exllamav2-0.1.7 with default dataset. I used a slightly modified quantization script to force use of highest bpw methods for all layers in the model (which is usually "1:8b_128g s4") to ensure max quality.

I also added a small fix in config file to set max default context at 128k as original Mistral-Nemo should have.

I tested this quant shortly in some random RPs (including ones over 8k context) and it seems to work fine.

Prompt Templates

Uses Mistral format.

Original readme below

This model is the miniature version of alpindale/magnum-72b-v1, a second entry in a series of models designed to replicate the prose quality of the Claude 3 models, specifically Sonnet and Opus. This model is fine-tuned on top of Mistral-Nemo-Base-2407. A new general purpose instruction dataset by kalomaze was added to the training mix for better coherence and general alignment. We are working on improving our dataset and training procedures, so expect new versions to come out soon.

Prompting

Model has been Instruct tuned with the Mistral formatting. A typical input would look like this:

"""[INST] Hi there! [/INST]Nice to meet you!</s>[INST] Can I ask a question? [/INST]
"""

Credits

This model has been a team effort, credits go to:

Sao10K and kalomaze for help with (and cleaning up!) the dataset.
alpindale for the training.
Various other people for their continued help as we tuned the parameters, restarted failed runs. In no particular order: Doctor Shotgun, Lucy, Nopm, Mango, Intervitens, and the rest of the Silly Tilly.

Safety

...