KingOmini_14b_v1

KingOmini_14b_v1 is a merge of the following models using mergekit:

• kainatq/KingOmini_14b_v1_M1
• v000000/Qwen2.5-14B-Gutenberg-1e-Delta
• clouditera/secgpt

🧩 Configuration

merge_method: dare_ties
base_model: Qwen/Qwen2.5-14B-Instruct  # Still use ORIGINAL base for task vectors

models:
  # Stage 1 result (treat as single model)
  - model: kainatq/KingOmini_14b_v1_M1
    parameters:
      weight: 1.0

  # Medium priority: Literary creativity
  - model: v000000/Qwen2.5-14B-Gutenberg-1e-Delta
    parameters:
      weight: 0.8

  # Lower priority: Security/coding (avoid over-restriction)
  - model: clouditera/secgpt
    parameters:
      weight: 0.5

parameters:
  density: 0.33
  int8_mask: true
  normalize: true```