mergekit-community/qwq-slerp-3

merge

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

Merge Details

Merge Method

This model was merged using the SLERP merge method.

Models Merged

The following models were included in the merge:

• Qwen/QwQ-32B
• Qwen/Qwen2.5-32B-Instruct

Configuration

The following YAML configuration was used to produce this model:

merge_method:        slerp
base_model:          Qwen/QwQ-32B
tokenizer_source:    Qwen/QwQ-32B
dtype:               bfloat16
out_dtype:           bfloat16
parameters:         
  int8_mask:         true
  normalize:         true
  rescale:           false
  t:
    - value:         0.50
slices:
  - sources:
      - model:       Qwen/QwQ-32B
        layer_range: [ 0, 8 ]
      - model:       Qwen/QwQ-32B
        layer_range: [ 0, 8 ]
  - sources:
      - model:       Qwen/Qwen2.5-32B-Instruct
        layer_range: [ 8, 16 ]
      - model:       Qwen/QwQ-32B
        layer_range: [ 8, 16 ]
  - sources:
      - model:       Qwen/Qwen2.5-32B-Instruct
        layer_range: [ 16, 24 ]
      - model:       Qwen/QwQ-32B
        layer_range: [ 16, 24 ]
  - sources:
      - model:       Qwen/Qwen2.5-32B-Instruct
        layer_range: [ 24, 32 ]
      - model:       Qwen/Qwen2.5-32B-Instruct
        layer_range: [ 24, 32 ]
  - sources:
      - model:       Qwen/Qwen2.5-32B-Instruct
        layer_range: [ 32, 40 ]
      - model:       Qwen/Qwen2.5-32B-Instruct
        layer_range: [ 32, 40 ]
  - sources:
      - model:       Qwen/QwQ-32B
        layer_range: [ 40, 48 ]
      - model:       Qwen/Qwen2.5-32B-Instruct
        layer_range: [ 40, 48 ]
  - sources:
      - model:       Qwen/QwQ-32B
        layer_range: [ 48, 56 ]
      - model:       Qwen/Qwen2.5-32B-Instruct
        layer_range: [ 48, 56 ]
  - sources:
      - model:       Qwen/QwQ-32B
        layer_range: [ 56, 64 ]
      - model:       Qwen/QwQ-32B
        layer_range: [ 56, 64 ]