A continuation of my Control series BUT ON NEMO! • 4 items • Updated
exl2 quant (measurement.json in main branch)
check revisions for quants
A finetune of Mistral-Nemo-Instruct-2407 using conversational data, aiming for prose that's best described as 'short' and 'sweet.' The model strictly adheres to one-on-one roleplay and is very dialogue heavy.
Quants
GGUF: https://huggingface.co/Delta-Vector/Ohashi-NeMo-12B-gguf
EXL2 : https://huggingface.co/Delta-Vector/Ohashi-NeMo-12B-EXL2
Prompting
Model has been tuned with the Mistral formatting. A typical input would look like this:
<s>[INST] SYSTEM MESSAGE
USER MESSAGE[/INST] ASSISTANT MESSAGE</s>[INST] USER MESSAGE[/INST]
System Prompting
I would highly recommend using either Euryale's system prompt or the EVA system prompt with the model.
Axolotl config
Credits
Thank you to Lucy Knada, Intervitens, Tav, Trappu, Cgato, Kubernetes Bad and the rest of Anthracite
Training
The training was done for 4 epochs. We used 4 x RTX 3090s GPUs graciously provided by Intervitens for the fine-tuning of the model.
Safety
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support