BlueNipples's picture
Update README.md
514c8a4 verified
|
raw
history blame
1.36 kB
---
license: apache-2.0
tags:
- Solar
- Mistral
- Roleplay
---
![image/png](https://cdn-uploads.huggingface.co/production/uploads/64bb1109aaccfd28b023bcec/bjOB_8BsqVteKKxARPc13.png)
### Summary
I did a more aggressive gradient SLERP on the base model with a medical lora added to Frostwind preserving more of the prose model and the prose seemed better without degredation in coherency. So then I redid the dares ties merge of it, with it's parent models, but with a slightly lower weight to FrostWind.
Prose is a bit stronger, coherency slightly weaker with this variant. It's excellent if you are prepared to regen. Really happy with both these models.
### Recipe
Formula for the DARE TIES merge is as follows (FrostMed is Frostwind with Solarmed at 0.15 weight):
- model: ./Frostmaid
parameters:
density: [0.45] # density gradient
weight: 0.23
- model: ./FrostMed
parameters:
density: [0.35] # density gradient
weight: 0.18
- model: ./SnowLotus-10.7B-v2
parameters:
density: [1] # density gradient
weight: 1
Resources used:
https://huggingface.co/BlueNipples/SnowLotus-v2-10.7B
https://huggingface.co/NyxKrage/FrostMaid-10.7B-TESTING-pt
https://huggingface.co/Sao10K/Frostwind-10.7B-v1
https://huggingface.co/NyxKrage/Solar-Doc-10.7B-Lora
https://github.com/cg123/mergekit/tree/main