mergekit-uploader's picture
Upload folder using huggingface_hub
acc1135 verified
---
base_model:
- MarinaraSpaghetti/NemoMix-Unleashed-12B
- IlyaGusev/saiga_nemo_12b
- TheDrummer/Rocinante-12B-v1.1
- Vikhrmodels/Vikhr-Nemo-12B-Instruct-R-21-09-24
library_name: transformers
tags:
- mergekit
- merge
---
# merge
This is a merge of pre-trained language models created using [mergekit](https://github.com/cg123/mergekit).
## Merge Details
### Merge Method
This model was merged using the [Linear DELLA](https://arxiv.org/abs/2406.11617) merge method using [IlyaGusev/saiga_nemo_12b](https://huggingface.co/IlyaGusev/saiga_nemo_12b) as a base.
### Models Merged
The following models were included in the merge:
* [MarinaraSpaghetti/NemoMix-Unleashed-12B](https://huggingface.co/MarinaraSpaghetti/NemoMix-Unleashed-12B)
* [TheDrummer/Rocinante-12B-v1.1](https://huggingface.co/TheDrummer/Rocinante-12B-v1.1)
* [Vikhrmodels/Vikhr-Nemo-12B-Instruct-R-21-09-24](https://huggingface.co/Vikhrmodels/Vikhr-Nemo-12B-Instruct-R-21-09-24)
### Configuration
The following YAML configuration was used to produce this model:
```yaml
models:
- model: IlyaGusev/saiga_nemo_12b
parameters:
weight: 0.5 # Основной акцент на русском языке
density: 0.4
- model: MarinaraSpaghetti/NemoMix-Unleashed-12B
parameters:
weight: 0.2 # РП модель, чуть меньший вес из-за ориентации на английский
density: 0.4
- model: TheDrummer/Rocinante-12B-v1.1
parameters:
weight: 0.2 # Увеличенный вес для усиления РП аспектов
density: 0.5 # Повышенная плотность для более сильного влияния
- model: Vikhrmodels/Vikhr-Nemo-12B-Instruct-R-21-09-24
parameters:
weight: 0.25 # Русскоязычная поддержка и баланс
density: 0.4
merge_method: della_linear
base_model: IlyaGusev/saiga_nemo_12b
parameters:
epsilon: 0.05
lambda: 1
dtype: float16
tokenizer_source: base
```