This is a merge of pre-trained language models created using
mergekit.
This model was merged using the
DARE TIES merge method using
deepseek-ai/DeepSeek-R1-Distill-Qwen-7B as a base.
1models:
2 - model: deepseek-ai/DeepSeek-R1-Distill-Qwen-7B
3 # Sem parâmetros = Base sólida (Referência)
4 - model: nvidia/OpenMath-Nemotron-7B
5 parameters:
6 density: 0.85
7 weight: 0.3
8
9merge_method: dare_ties
10base_model: deepseek-ai/DeepSeek-R1-Distill-Qwen-7B
11parameters:
12 int8_mask: true
13dtype: bfloat16
14
