win10/karcher-max-iter1000-32b

I am the creator and developer of the karcher method. If you want to cooperate, please contact my email.

Please support my ko-fi. https://ko-fi.com/ogodwin10

merge

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

Merge Details

Merge Method

This model was merged using the Karcher Mean merge method using Qwen/Qwen2.5-Coder-32B-Instruct as a base.

Models Merged

The following models were included in the merge:

• Skywork/Skywork-OR1-32B-Preview
• nvidia/OpenMath-Nemotron-32B
• EVA-UNIT-01/EVA-Qwen2.5-32B-v0.2
• all-hands/openhands-lm-32b-v0.1
• internlm/OREAL-32B
• Qwen/QwQ-32B

Configuration

The following YAML configuration was used to produce this model:

models:
  - model: all-hands/openhands-lm-32b-v0.1
  - model: internlm/OREAL-32B
  - model: nvidia/OpenMath-Nemotron-32B
  - model: Qwen/QwQ-32B
  - model: EVA-UNIT-01/EVA-Qwen2.5-32B-v0.2
  - model: Qwen/Qwen2.5-Coder-32B-Instruct
  - model: Skywork/Skywork-OR1-32B-Preview
merge_method: karcher
base_model: Qwen/Qwen2.5-Coder-32B-Instruct
parameters:
  max_iter: 1000
normalize: true
int8_mask: true
tokenizer_source: base
dtype: bfloat16