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 ./BaseConfigA as a base.
The following models were included in the merge:
The following YAML configuration was used to produce this model:
#Base Config A - Config C
models:
- model: TheDrummer/Magidonia-24B-v4.3
- model: ApocalypseParty/Magi-24B-SFT-v4-1-DPO-2
- model: Delta-Vector/Rei-24B-KTO
- model: Delta-Vector/MS3.2-Austral-Winton
select_topk: .15
merge_method: sce
base_model: ./BaseConfigA
out_dtype: bfloat16
dtype: float32
tokenizer:
source: base