FuseR1-QwQ-R1-32B / fusechat-sce.yml
AALF's picture
Upload folder using huggingface_hub
481a090 verified
models:
- model: Qwen/Qwen2.5-32B
- model: deepseek-ai/DeepSeek-R1-Distill-Qwen-32B
- model: Qwen/QwQ-32B
merge_method: sce
base_model: Qwen/Qwen2.5-32B
parameters:
select_topk: 1.0
dtype: bfloat16