mergekitty

mergekitty is a toolkit for merging pre-trained language models. It uses an out-of-core approach so you can run surprisingly complex merges on modest hardware — entirely on CPU, or with as little as 8 GB of VRAM.

What's this fork?

Forked from mergekit (originally by Charles Goddard, then maintained by Arcee.ai). The original project switched to a BSL license after a ton of community contribution, then switched back to LGPL but added a CLA that lets them relicense at will. So here we are.

Why merge models?

Model merging is chaos magick. Done right, the result is better than any of its inputs. It's been proven repeatedly and nobody fully understands why. Ship it.

Features

• Works with Llama 3, Qwen 3 (Dense & MoE), Mistral, GLM4, GPT-NeoX, BERT, and more
• Tons of merge methods — arguably too many
• GPU or CPU — your call
• Lazy tensor loading for low memory use
• Interpolated gradient parameters for fine control
• Layer-stacking / "Frankenmerging" (à la Goliath, Midnight Miqu)
• MoE merging and LoRA extraction

Install

# recommended — isolated tool install
uv tool install mergekitty

# or just pip
pip install mergekitty

# from source
git clone https://github.com/allura-org/mergekitty.git
cd mergekitty
pip install -e .

Usage

mergekitty-yaml path/to/config.yml ./output-model [--compute-device cuda] [--storage-device cuda] [--load-to-compute] [--lazy-unpickle] [--allow-crimes]

Run mergekitty-yaml --help for the full list of options.

Sharing on Huggingface

mergekitty generates a README.md for your merge. Edit it, keep it as-is, whatever — then upload:

huggingface-cli login
huggingface-cli upload your_username/my-cool-model ./output-model .

Merge Configuration

Configs are YAML. The main fields:

Field	Description
merge_method	Which algorithm to use (see below)
slices / models	Input model definitions (mutually exclusive)
base_model	Base model, for methods that need one
parameters	Weights, densities, etc. — specifiable at multiple levels
dtype	Data type for the merge
tokenizer	Vocabulary and embedding configuration
chat_template	Override the output chat template

Parameters

Parameters (weight, density, etc.) can be set at four levels, most-specific wins:

1. slices.*.sources.parameters — per input slice
2. slices.*.parameters — per output slice
3. models.*.parameters — per input model
4. parameters — global fallback

Values can be scalars or interpolated gradients (a list of floats for smooth transitions across layers).

Tokenizer

Use the tokenizer field for full control, or tokenizer_source for the simple legacy behavior.

tokenizer:
  source: union          # "union", "base", or a model path
  tokens:                # optional: per-token embedding overrides
    :
      source: "chatml_model"
    <|start_header_id|>:
      source: "llama3_model"
      force: true
  pad_to_multiple_of: null

Defaults are sensible: base model embeddings win if the token exists there, single-model tokens use that model, otherwise it averages. You can override any of this per-token.

Chat Template

chat_template: "auto"    # picks the most common template from inputs
# or: "alpaca", "chatml", "llama3", "mistral", "exaone"
# or: a raw Jinja2 template string

Examples

Check examples/ for real configs.

Merge Methods

Method	merge_method	Multi-Model	Needs Base
Linear (Model Soups)	linear	✅	❌
SLERP	slerp	✅*	✅
Nearswap	nearswap	❌	✅
Task Arithmetic	task_arithmetic	✅	✅
TIES	ties	✅	✅
DARE + TIES	dare_ties	✅	✅
DARE + Linear	dare_linear	✅	✅
Passthrough	passthrough	❌	❌
Model Breadcrumbs	breadcrumbs	✅	✅
Breadcrumbs + TIES	breadcrumbs_ties	✅	✅
Model Stock	model_stock	✅	✅
DELLA	della	✅	✅
DELLA + Linear	della_linear	✅	✅
SCE	sce	✅	✅

* SLERP supports two to three models.

Linear

Weighted average. Simple, classic, effective.

• weight — relative weighting per tensor
• normalize — normalize weights across models (default: true)

SLERP

Spherical interpolation. Supports t (classic SLERP, 0 = base, 1 = other) or weight (NuSLERP-style per-tensor weighting).

• nuslerp_flatten — treat tensor as flat vector vs. row/column-wise
• nuslerp_row_wise — SLERP row vectors instead of column vectors

Nearswap

Interpolates between base and secondary model when similarity drops below threshold t.

Task Arithmetic

Subtract base model → get "task vectors" → merge them linearly → add base back. Great for models fine-tuned from a common ancestor. Also the mental model behind most of the fancier methods.

TIES

Task arithmetic + sparsification + sign consensus. Lets you merge more models without them stepping on each other.

• density — fraction of task vector weights to keep

DARE

Random pruning with rescaling, instead of TIES's magnitude-based sparsification. Works with TIES sign consensus (dare_ties) or without (dare_linear).

Passthrough

No-op. Passes tensors through unchanged. Useful for layer-stacking / frankenmerging where you only have one input per slice.

Model Breadcrumbs

Drops both tiny and huge differences from base. Works with (breadcrumbs_ties) or without (breadcrumbs) TIES.

• density — fraction of weights to keep
• gamma — fraction of largest-magnitude differences to remove (paper's β)
• Defaults: density: 0.9, gamma: 0.01

Model Stock

Geometric trick to compute good linear weights. Needs at least three models including a base.

DELLA

Adaptive pruning based on magnitude ranking — keeps important changes, drops the rest. Like DARE but smarter about what it prunes.

• density — fraction of weights to keep
• epsilon — spread of drop probabilities (range: density ± epsilon)
• lambda — scaling factor for merged deltas

SCE

Selects high-variance elements, computes matrix-level weights, erases minority contributions.

• select_topk — fraction of high-variance elements to retain

LoRA Extraction

Extract PEFT-compatible LoRA adapters from finetuned models:

mergekitty-extract-lora finetuned_model base_model output_path --rank=32

MoE Merging

Merge dense models into a Mixture of Experts with mergekitty-moe. See the MoE docs.

Development

Uses Hatch + uv:

uv tool install hatch
hatch test              # run tests
hatch run lint          # ruff linting
hatch run format        # ruff formatting
hatch run mergekitty-yaml examples/bio-merge.yml ./bio-merge --compute-device cuda

Citation

If you use mergekitty in research, please cite the original mergekit paper:

@inproceedings{goddard-etal-2024-arcees,
    title = "Arcee{'}s {M}erge{K}it: A Toolkit for Merging Large Language Models",
    author = "Goddard, Charles  and
      Siriwardhana, Shamane  and
      Ehghaghi, Malikeh  and
      Meyers, Luke  and
      Karpukhin, Vladimir  and
      Benedict, Brian  and
      McQuade, Mark  and
      Solawetz, Jacob",
    booktitle = "Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing: Industry Track",
    month = nov,
    year = "2024",
    pages = "477--485",
    url = "https://aclanthology.org/2024.emnlp-industry.36",
}