Sela223/asp-9b

output

This is a merge of pre-trained language models created using mergekit.

Merge Details

Merge Method

This model was merged using the SpecLERP v4 merge method using ReadyArt/Omega-Evolution-9B-v2.0 as a base.

Models Merged

The following models were included in the merge:

• Dxniz/NaNovel-9B

Configuration

The following YAML configuration was used to produce this model:

merge_method: splerp_v4
base_model: ReadyArt/Omega-Evolution-9B-v2.0
dtype: bfloat16
tokenizer_source: base
architecture: qwen3_5
modules:
  text_decoder:
    architecture: qwen3_5_text
    models:
      - model: ReadyArt/Omega-Evolution-9B-v2.0
      - model: Dxniz/NaNovel-9B
    parameters:
      equal_weight: true
      t:
        - filter: ".*self_attn\\.q_proj.*"
          value: 0.5
        - filter: ".*self_attn\\.k_proj.*"
          value: 0.5
        - filter: ".*self_attn\\.v_proj.*"
          value: 0.5
        - filter: ".*self_attn\\.o_proj.*"
          value: 0.5
        - filter: ".*linear_attn.*"
          value: 0.5
        - filter: ".*(gate_proj|up_proj|down_proj).*"
          value: 0.5
        - value: 0.5
  vision_tower:
    architecture: qwen3_5_vision
    models:
      - model: ReadyArt/Omega-Evolution-9B-v2.0
      - model: Dxniz/NaNovel-9B
    parameters:
      t: 0.5