legion

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

Merge Details

Merge Method

This model was merged using the DARE TIES merge method using kromcomp/L3.1-Stygianv1-12B as a base.

Models Merged

The following models were included in the merge:

Configuration

The following YAML configuration was used to produce this model:

base_model: kromcomp/L3.1-Stygianv1-12B
dtype: float32
merge_method: dare_ties
parameters:
  normalize: 1.0
slices:
- sources:
  - layer_range: [0, 50]
    model: kromcomp/L3.1-Beastv1-12B
    parameters:
      density:
      - filter: self_attn
        value: 0.7
      - filter: mlp
        value: 0.5
      - value: 0.6
      lambda:
      - filter: self_attn
        value: 1.05
      - value: 1.0
      weight: 0.5
  - layer_range: [0, 50]
    model: kromcomp/L3.1-Stygianv1-12B
    parameters:
      density:
      - filter: self_attn
        value: 0.5
      - filter: mlp
        value: 0.7
      - value: 0.6
      lambda: 1.0
      weight: 0.5
tokenizer:
  source: kromcomp/L3.1-Beastv1-12B
Downloads last month
3
Safetensors
Model size
12B params
Tensor type
F32
ยท
Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐Ÿ™‹ Ask for provider support

Model tree for kromcomp/L3.1-Legionv2-12B