ThijsL202/EVA-Hamanasu-Magnum-QwQ-32B

merge

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

I created it because I like win10/EVA-QwQ-32B-Preview a lot, sure, it can be repetitive and it's not really affected by the sampler settings when using the default koboldcpp order, but it's still one of the best for the way I roleplay.

I might try this again using different models (Instead of EVA-UNIT-01/EVA-Qwen2.5-32B-v0.2) with the same merge method, any recommendations? Feel free to ask me.

Merge Details

Merge Method

This model was merged using the TIES merge method using Delta-Vector/Hamanasu-Magnum-QwQ-32B as a base.

Models Merged

The following models were included in the merge:

• EVA-UNIT-01/EVA-Qwen2.5-32B-v0.2

Configuration

The following YAML configuration was used to produce this model:

models:
  - model: Delta-Vector/Hamanasu-Magnum-QwQ-32B
    parameters:
      density: 1
      weight: 1 
      
  - model: EVA-UNIT-01/EVA-Qwen2.5-32B-v0.2
    parameters:
      density: 1
      weight: 1 
      
  - model: Delta-Vector/Hamanasu-Magnum-QwQ-32B
    parameters:
      density: 1
      weight: 1 
      
  - model: EVA-UNIT-01/EVA-Qwen2.5-32B-v0.2
    parameters:
      density: 1
      weight: 1 

merge_method: ties
base_model: Delta-Vector/Hamanasu-Magnum-QwQ-32B
parameters:
  normalize: false
  int8_mask: true
dtype: bfloat16