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 Undi95/MXLewd-L2-20B as a base.
The following models were included in the merge:
The following YAML configuration was used to produce this model:
merge_method: sce
models:
- model: Undi95/PsyMedRP-v1-20B
- model: Elfrino/PsyMedLewd_v4
- model: Elfrino/PsychoPomp-20B
base_model: Undi95/MXLewd-L2-20B
tokenizer:
source: base
parameters:
select_topk: 0.693
dtype: float32
out_dtype: bfloat16
normalize: true