This is a merge of pre-trained language models created using
mergekit.
This model was merged using the
SLERP merge method.
1slices:
2 - sources:
3 - model: Qwen/Qwen2-1.5B
4 layer_range: [0, 28] # Covering all 28 layers (0 to 27)
5 - model: jinaai/ReaderLM-v2
6 layer_range: [0, 28] # Ensure this model has a similar layer structure
7merge_method: slerp
8base_model: Qwen/Qwen2-1.5B
9parameters:
10 t:
11 - filter: self_attn
12 value: [0, 0.5, 0.3, 0.7, 1]
13 - filter: mlp
14 value: [1, 0.5, 0.7, 0.3, 0]
15 - value: 0.5 # Fallback for any tensors not covered by the filters
16dtype: bfloat16 # Or float16, depending on your preference and model dtype
17
18
