mergekit-community/SuperQwen-2.5-1.5B

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/Qwen2.5-1.5B-Instruct as a base.

Models Merged

The following models were included in the merge:

• Qwen/Qwen2.5-Coder-1.5B-Instruct
• Qwen/Qwen2.5-Math-1.5B-Instruct

Configuration

The following YAML configuration was used to produce this model:

merge_method: ties  # Use TIES for merging multiple models
base_model: Qwen/Qwen2.5-1.5B-Instruct  # Base model for the merge
dtype: bfloat16  # Data type for the merged model

models:
  - model: Qwen/Qwen2.5-1.5B-Instruct  # Base model
    parameters:
      weight: 0.5  # Weight for the base model

  - model: Qwen/Qwen2.5-Math-1.5B-Instruct  # Math-focused model
    parameters:
      density: 0.6  # Retain 60% of significant parameters
      weight: 0.3  # Weight for the math model

  - model: Qwen/Qwen2.5-Coder-1.5B-Instruct  # Code-focused model
    parameters:
      density: 0.6  # Retain 60% of significant parameters
      weight: 0.2  # Weight for the coder model

parameters:
  normalize: true  # Normalize weights to ensure compatibility
  int8_mask: true  # Optimize memory and computational efficiency