---
base_model:
- Theros/Q2.5-ColdBrew-R1-Obsidian
- SvalTek/Q2.5-ColdBrew-R1-Indigo
- Theros/Q2.5-DeepSeek-R1-DeepThink-test1
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 [Model Stock](https://arxiv.org/abs/2403.19522) merge method using [SvalTek/Q2.5-ColdBrew-R1-Indigo](https://huggingface.co/SvalTek/Q2.5-ColdBrew-R1-Indigo) as a base.

### Models Merged

The following models were included in the merge:
* [Theros/Q2.5-ColdBrew-R1-Obsidian](https://huggingface.co/Theros/Q2.5-ColdBrew-R1-Obsidian)
* [Theros/Q2.5-DeepSeek-R1-DeepThink-test1](https://huggingface.co/Theros/Q2.5-DeepSeek-R1-DeepThink-test1)

### Configuration

The following YAML configuration was used to produce this model:

```yaml
name: SvalTek/Q2.5-ColdBrew-R1-Atrax
models:
  - model: SvalTek/Q2.5-ColdBrew-R1-Indigo
  - model: Theros/Q2.5-DeepSeek-R1-DeepThink-test1
  - model: Theros/Q2.5-ColdBrew-R1-Obsidian
    parameters:
      weight: 0.3
merge_method: model_stock
base_model: SvalTek/Q2.5-ColdBrew-R1-Indigo
normalize: true
int8_mask: true
dtype: bfloat16
tokenizer_source: SvalTek/Q2.5-ColdBrew-R1-Indigo
```