BenevolenceMessiah/QwQ-32B-Preview-Self-instruct-4x-TIES-MGS-v1.0

merge

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

Merge Details

Merge Method

This model was merged using the TIES merge method using Qwen/QwQ-32B-Preview as a base.

Models Merged

The following models were included in the merge:

• fblgit/TheBeagle-v2beta-32B-MGS
• EVA-UNIT-01/EVA-Qwen2.5-32B-v0.2
• rombodawg/Rombos-Coder-V2.5-Qwen-32b
• rombodawg/Rombos-LLM-V2.5-Qwen-32b

Configuration

The following YAML configuration was used to produce this model:

# QwQ-32B-Preview-Self-instruct-4x-TIES-MGS-v1.0

models:
  - model: rombodawg/Rombos-LLM-V2.5-Qwen-32b # Self-instruct fine-tuning on Qwen2.5
    parameters:
      density: 1.0
      weight: 1.0
  - model: EVA-UNIT-01/EVA-Qwen2.5-32B-v0.2 # RP/storywriting specialist model (hopefully correcting some of the known limitations of the base model)
    parameters:
      density: 1.0
      weight: 1.0
  - model: fblgit/TheBeagle-v2beta-32B-MGS # MGS training
    parameters:
      density: 1.0
      weight: 1.0
  - model: rombodawg/Rombos-Coder-V2.5-Qwen-32b # Self-instruct fine-tuning on Qwen2.5-Coder
    parameters:
      density: 1.0
      weight: 1.0

merge_method: ties
base_model: Qwen/QwQ-32B-Preview # Advanced AI reasoning capabilities
parameters:
  normalize: true
  int8_mask: false
dtype: bfloat16
tokenizer_source: union