arcee-ai / mergekit
Tools for merging pretrained large language models.
Why this job: works in the areas we cover, but does not say which job it does
What it does
- Supports Llama, Mistral, GPT-NeoX, StableLM, and more
- GPU or CPU execution
- Lazy loading of tensors for low memory use
- Interpolated gradients for parameter values (inspired by Gryphe's BlockMerge_Gradient script)
Get started
git clone https://github.com/arcee-ai/mergekit.git
cd mergekit
pip install -e .
# install the package and make scripts available - Maintenance
- Active
- Last commit 12 days ago
- Stars
- 7,448
- +13 this week
- Licence
- LGPL-3.0
- Share-alike for changes
- Downloads for
- None listed
- No release files
- Latest version
- v0.1.5
- Released 23 days ago
- Built with
- Python
- 99% of the code
Is it looked after?
- Contributors
- 8
- Last 90 days
- Issues
- 45% closed
- 239 still open
- Pull requests
- 190 merged
- 38 waiting
- Releases
- 3
- 1 in the last year
- Contributing guide (opens in new tab)
- No security policy
- No code of conduct
More: languages, topics and project record
Languages
- Python 99%
- Jinja <1%
- Jupyter Notebook <1%
Topics
- llama
- llm
- model-merging
Project record
- JobOther AI tools
- StatusListed
- Commits316
- Forks803
- Watching60
- Stars in 28 days+56
- CreatedAug 21, 2023
- First seen by usNov 7, 2026
- CheckedNov 7, 2026