FuseChat: Knowledge Fusion of Chat Models
Paper
•
2408.07990
•
Published
•
14
This is a merge of pre-trained language models created using mergekit.
This model was merged using the SCE merge method using Qwen/Qwen2.5-14B as a base.
The following models were included in the merge:
The following YAML configuration was used to produce this model:
models:
# Pivot model
- model: Qwen/Qwen2.5-14B
# Target models
- model: Qwen/Qwen2.5-Coder-14B
- model: deepseek-ai/DeepSeek-R1-Distill-Qwen-14B
merge_method: sce
base_model: Qwen/Qwen2.5-14B
parameters:
select_topk: 1.0
dtype: bfloat16