Tarek07's picture
Add files using upload-large-folder tool
f4643f7 verified
---
base_model:
- TareksLab/M-MERGE4
- TareksLab/M-MERGE3
- TareksLab/M-MERGE2
- TareksLab/M-BASE-SCE
- TareksLab/M-MERGE1
library_name: transformers
tags:
- mergekit
- merge
---
# LegionDL
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 [TareksLab/M-BASE-SCE](https://huggingface.co/TareksLab/M-BASE-SCE) as a base.
### Models Merged
The following models were included in the merge:
* [TareksLab/M-MERGE4](https://huggingface.co/TareksLab/M-MERGE4)
* [TareksLab/M-MERGE3](https://huggingface.co/TareksLab/M-MERGE3)
* [TareksLab/M-MERGE2](https://huggingface.co/TareksLab/M-MERGE2)
* [TareksLab/M-MERGE1](https://huggingface.co/TareksLab/M-MERGE1)
### Configuration
The following YAML configuration was used to produce this model:
```yaml
models:
- model: TareksLab/M-MERGE4
parameters:
weight: 0.15
density: 0.7
epsilon: 0.2
lambda: 1.1
- model: TareksLab/M-MERGE3
parameters:
weight: 0.20
density: 0.7
epsilon: 0.2
lambda: 1.1
- model: TareksLab/M-MERGE2
parameters:
weight: 0.20
density: 0.7
epsilon: 0.2
lambda: 1.1
- model: TareksLab/M-MERGE1
parameters:
weight: 0.25
density: 0.7
epsilon: 0.2
lambda: 1.1
- model: TareksLab/M-BASE-SCE
parameters:
weight: 0.20
density: 0.7
epsilon: 0.1
lambda: 1.0
merge_method: della_linear
base_model: TareksLab/M-BASE-SCE
parameters:
normalize: false
int8_mask: true
dtype: bfloat16
chat_template: llama3
tokenizer:
source: TareksLab/M-TOKENIZER-SCE
```