This is a merge of pre-trained language models created using
mergekit.
If you enjoy this model, don't forget to leave feedback and a like on this model page.
This model was merged using the
Reinforced Agent Merging Plus (Tensor-Local) merge method using
Qwen/Qwen3.8-27B as a base.
1base_model: Qwen/Qwen3.8-27B
2dtype: bfloat16
3merge_method: ramplus_tl
4modules:
5 default:
6 slices:
7 - sources:
8 - layer_range: [0, 0]
9 model: DavidAU/Qwen3.8-27B-Cold-Fusion-GAIN-V1.1
10 - layer_range: [0, 0]
11 model: beyoru/Kiwen1.1-27B
12 - layer_range: [0, 0]
13 model: TeichAI/Qwen3.8-27B-Fable-Distill
14 - layer_range: [0, 0]
15 model: Qwen/Qwen3.8-27B
16 model.language_model.layers:
17 slices:
18 - sources:
19 - layer_range: [0, 64]
20 model: DavidAU/Qwen3.8-27B-Cold-Fusion-GAIN-V1.1
21 - layer_range: [0, 64]
22 model: beyoru/Kiwen1.1-27B
23 - layer_range: [0, 64]
24 model: TeichAI/Qwen3.8-27B-Fable-Distill
25 - layer_range: [0, 64]
26 model: Qwen/Qwen3.8-27B
27 model.visual.blocks:
28 slices:
29 - sources:
30 - layer_range: [0, 27]
31 model: DavidAU/Qwen3.8-27B-Cold-Fusion-GAIN-V1.1
32 - layer_range: [0, 27]
33 model: beyoru/Kiwen1.1-27B
34 - layer_range: [0, 27]
35 model: TeichAI/Qwen3.8-27B-Fable-Distill
36 - layer_range: [0, 27]
37 model: Qwen/Qwen3.8-27B
38out_dtype: bfloat16
39parameters:
40 r: 0.25
