models: | |
- model: djuna/Q2.5-Veltha-14B | |
- model: huihui-ai/DeepSeek-R1-Distill-Qwen-14B-abliterated-v2 | |
merge_method: slerp | |
base_model: djuna/Q2.5-Veltha-14B | |
parameters: | |
t: | |
- filter: self_attn | |
value: [0.2, 0.25, 0.3, 0.25, 0.2] | |
- filter: "q_proj|k_proj|v_proj" | |
value: [0.2, 0.25, 0.3, 0.25, 0.2] | |
- filter: "up_proj|down_proj" | |
value: [0.2, 0.3, 0.4, 0.3, 0.2] | |
- filter: mlp | |
value: [0.25, 0.35, 0.55, 0.35, 0.25] | |
- value: 0.45 # default for other components | |
dtype: bfloat16 | |