Model-merging-slerp / mergekit_config.yml
sshweta3's picture
Upload folder using huggingface_hub
1d1e601 verified
raw
history blame contribute delete
394 Bytes
slices:
- sources:
- model: Qwen/Qwen2.5-32B
layer_range: [0, 40]
- model: Qwen/Qwen2.5-32B-Instruct
layer_range: [0, 40]
merge_method: slerp
base_model: Qwen/Qwen2.5-32B
parameters:
t:
- filter: self_attn
value: [0, 0.5, 0.3, 0.7, 1]
- filter: mlp
value: [1, 0.5, 0.7, 0.3, 0]
- value: 0.5 # fallback for rest of tensors
dtype: bfloat16