merged-reasoner-coder

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

Merge Details

Merge Method

This model was merged using the DARE TIES merge method using Qwen/Qwen2.5-Coder-7B-Instruct as a base.

Models Merged

The following models were included in the merge:

• open-thoughts/OpenThinker3-7B

Configuration

The following YAML configuration was used to produce this model:

merge_method: dare_ties
base_model: Qwen/Qwen2.5-Coder-7B-Instruct
models:
  - model: open-thoughts/OpenThinker3-7B
    parameters:
      density: 0.5
      weight:
        - filter: embed_tokens
          value: 0
        - filter: lm_head
          value: 0
        - value: 0.5
parameters:
  normalize: false   # keep our explicit weight instead of rescaling to sum-to-1
dtype: bfloat16
tokenizer_source: Qwen/Qwen2.5-Coder-7B-Instruct   # keep Coder's ChatML template + tool-call special tokens