sometimesanotion/Base-Qwenvergence

merge

This is nothing exciting. It's just part of an experiment.

Merge Details

Merge Method

This model was merged using the DARE TIES merge method using Qwen/Qwen2.5-14B as a base.

Models Merged

The following models were included in the merge:

• sometimesanotion/Abliterate-Qwenvergence

Configuration

The following YAML configuration was used to produce this model:

name:                Base-Qwenvergence-dare_ties
merge_method:        dare_ties
base_model:          Qwen/Qwen2.5-14B
tokenizer_source:    base
dtype:               float32
out_dtype:           bfloat16
parameters:         
  density:           1.00
  weight:            1.00
  int8_mask:         true
  normalize:         true
  rescale:           false
slices:
  - sources:
    - { layer_range: [  0,  48 ], parameters: { density: 1.00, weight: 1.00 }, model: Qwen/Qwen2.5-14B }
    - { layer_range: [  0,  48 ], parameters: { density: 0.00, weight: 0.00 }, model: sometimesanotion/Abliterate-Qwenvergence }