|
--- |
|
base_model: |
|
- TheDrummer/Cydonia-22B-v1.1 |
|
- TheDrummer/Cydonia-22B-v1.2 |
|
- anthracite-org/magnum-v4-22b |
|
- TheDrummer/Cydonia-22B-v1.3 |
|
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 [DARE](https://arxiv.org/abs/2311.03099) [TIES](https://arxiv.org/abs/2306.01708) merge method using [anthracite-org/magnum-v4-22b](https://huggingface.co/anthracite-org/magnum-v4-22b) as a base. |
|
|
|
### Models Merged |
|
|
|
The following models were included in the merge: |
|
* [TheDrummer/Cydonia-22B-v1.1](https://huggingface.co/TheDrummer/Cydonia-22B-v1.1) |
|
* [TheDrummer/Cydonia-22B-v1.2](https://huggingface.co/TheDrummer/Cydonia-22B-v1.2) |
|
* [TheDrummer/Cydonia-22B-v1.3](https://huggingface.co/TheDrummer/Cydonia-22B-v1.3) |
|
|
|
### Configuration |
|
|
|
The following YAML configuration was used to produce this model: |
|
|
|
```yaml |
|
models: |
|
- model: anthracite-org/magnum-v4-22b |
|
parameters: |
|
weight: 1.0 # Magnum model to retain style and best writing |
|
- model: TheDrummer/Cydonia-22B-v1.3 |
|
parameters: |
|
weight: 0.4 # Cydonia v1.3 for creative/adventurous influence |
|
density: 0.9 # Retain more creative influence, but balance it |
|
- model: TheDrummer/Cydonia-22B-v1.2 |
|
parameters: |
|
weight: 0.3 # Cydonia v1.2 to add more creativity |
|
density: 0.85 # Slightly less density for extra creativity |
|
- model: TheDrummer/Cydonia-22B-v1.1 |
|
parameters: |
|
weight: 0.4 # Cydonia v1.1 for accuracy and evil/trolling thoughts |
|
density: 0.86 # Higher density for accuracy retention |
|
|
|
merge_method: dare_ties # Using dare_ties for better handling of interference |
|
base_model: anthracite-org/magnum-v4-22b |
|
parameters: |
|
density: 0.9 # General density for the overall merge |
|
epsilon: 0.1 # Maximum change in drop probabilities |
|
lambda: 1.2 # Scaling factor for final merged deltas |
|
dtype: bfloat16 |
|
|
|
``` |
|
|
