Bernstein-120b / README.md
bcse's picture
Update README.md
eae1fe6 verified
|
raw
history blame
2.06 kB
---
base_model:
- ShinojiResearch/Senku-70B-Full
- Sao10K/Euryale-1.3-L2-70B
library_name: transformers
tags:
- mergekit
- merge
---
# Bernstein-120b
This is a merge of pre-trained language models created using [mergekit](https://github.com/cg123/mergekit).
## Quants
* [mradermacher/Bernstein-120b-GGUF](https://huggingface.co/mradermacher/Bernstein-120b-GGUF)
* [mradermacher/Bernstein-120b-i1-GGUF](https://huggingface.co/mradermacher/Bernstein-120b-i1-GGUF)
## Merge Details
### Merge Method
This model was merged using the [linear](https://arxiv.org/abs/2203.05482) merge method.
### Models Merged
The following models were included in the merge:
* [ShinojiResearch/Senku-70B-Full](https://huggingface.co/ShinojiResearch/Senku-70B-Full)
* [Sao10K/Euryale-1.3-L2-70B](https://huggingface.co/Sao10K/Euryale-1.3-L2-70B)
### Configuration
The following YAML configuration was used to produce this model:
```yaml
merge_method: linear
parameters:
weight: 1.0
slices:
- sources:
- model: ShinojiResearch/Senku-70B-Full
layer_range: [0, 1]
- model: Sao10K/Euryale-1.3-L2-70B
layer_range: [0, 1]
parameters:
weight: 0
- sources:
- model: ShinojiResearch/Senku-70B-Full
layer_range: [1, 20]
- sources:
- model: Sao10K/Euryale-1.3-L2-70B
layer_range: [10, 30]
- sources:
- model: ShinojiResearch/Senku-70B-Full
layer_range: [20, 40]
- sources:
- model: Sao10K/Euryale-1.3-L2-70B
layer_range: [30, 50]
- sources:
- model: ShinojiResearch/Senku-70B-Full
layer_range: [40, 60]
- sources:
- model: Sao10K/Euryale-1.3-L2-70B
layer_range: [50, 70]
- sources:
- model: ShinojiResearch/Senku-70B-Full
layer_range: [60, 79]
- sources:
- model: ShinojiResearch/Senku-70B-Full
layer_range: [79, 80]
- model: Sao10K/Euryale-1.3-L2-70B
layer_range: [79, 80]
parameters:
weight: 0
dtype: float16
tokenizer_source: model:ShinojiResearch/Senku-70B-Full
```