Model Stock: All we need is just a few fine-tuned models
Paper
•
2403.19522
•
Published
•
13
This is a merge of pre-trained language models created using mergekit.
This model was merged using the Model Stock merge method using Theros/Qwen2.5-ColdBrew-R1-test as a base.
The following models were included in the merge:
The following YAML configuration was used to produce this model:
models:
- model: Theros/Qwen2.5-ColdBrew-R1-test2
- model: Theros/Qwen2.5-CreativeMix-7B
- model: deepseek-ai/DeepSeek-R1-Distill-Qwen-7B
parameters:
weight: 0.4
base_model: Theros/Qwen2.5-ColdBrew-R1-test
merge_method: model_stock
parameters:
dtype: bfloat16
tokenizer_source: Theros/Qwen2.5-ColdBrew-R1-test2