|
--- |
|
base_model: |
|
- NewEden/nemo-erebus |
|
- nbeerbower/mistral-nemo-gutenberg-12B-v4 |
|
- grimjim/mistralai-Mistral-Nemo-Instruct-2407 |
|
- nbeerbower/mistral-nemo-bophades-12B |
|
- TheDrummer/UnslopNemo-12B-v4.1 |
|
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 [task arithmetic](https://arxiv.org/abs/2212.04089) merge method using [NewEden/nemo-erebus](https://huggingface.co/NewEden/nemo-erebus) as a base. |
|
|
|
### Models Merged |
|
|
|
The following models were included in the merge: |
|
* [nbeerbower/mistral-nemo-gutenberg-12B-v4](https://huggingface.co/nbeerbower/mistral-nemo-gutenberg-12B-v4) |
|
* [grimjim/mistralai-Mistral-Nemo-Instruct-2407](https://huggingface.co/grimjim/mistralai-Mistral-Nemo-Instruct-2407) |
|
* [nbeerbower/mistral-nemo-bophades-12B](https://huggingface.co/nbeerbower/mistral-nemo-bophades-12B) |
|
* [TheDrummer/UnslopNemo-12B-v4.1](https://huggingface.co/TheDrummer/UnslopNemo-12B-v4.1) |
|
|
|
### Configuration |
|
|
|
The following YAML configuration was used to produce this model: |
|
|
|
```yaml |
|
models: |
|
- model: grimjim/mistralai-Mistral-Nemo-Instruct-2407 |
|
parameters: |
|
density: 0.2 |
|
weight: 0.23 |
|
- model: nbeerbower/mistral-nemo-bophades-12B |
|
parameters: |
|
density: 0.2 |
|
weight: 0.43 |
|
- model: nbeerbower/mistral-nemo-gutenberg-12B-v4 |
|
parameters: |
|
density: 0.2 |
|
weight: 0.43 |
|
- model: TheDrummer/UnslopNemo-12B-v4.1 |
|
parameters: |
|
density: 0.5 |
|
weight: 0.63 |
|
|
|
merge_method: task_arithmetic |
|
base_model: NewEden/nemo-erebus |
|
parameters: |
|
normalize: false |
|
int8_mask: true |
|
dtype: bfloat16 |
|
|
|
layer_parameters: |
|
- range: [0, 10] |
|
density_multiplier: 1.2 |
|
- range: [10, 20] |
|
density_multiplier: 1.0 |
|
- range: [20, 30] |
|
density_multiplier: 0.8 |
|
- range: [30, 40] |
|
density_multiplier: 0.6 |
|
|
|
regularization: |
|
- method: gradient_penalty |
|
scale: 0.05 |
|
- method: weight_clipping |
|
clip_range: [-0.2, 0.2] |
|
- method: random_noise |
|
scale: 0.01 |
|
- method: attention_dropout |
|
scale: 0.1 |
|
|
|
postprocessing: |
|
- operation: entropy_regularization |
|
scale: 0.05 |
|
- operation: non_linear_scaling |
|
parameters: |
|
function: tanh |
|
- operation: sharpening |
|
intensity: 0.5 |
|
- operation: gaussian_smoothing |
|
sigma: 1.5 |
|
- operation: normalize |
|
- operation: dynamic_scaling |
|
scale_range: [0.8, 1.2] |
|
- operation: smoothing |
|
parameters: |
|
adaptive: true |
|
range: [0.85, 1.15] |
|
kernel_size: 5 |
|
``` |
|
|
