LLM Model Merging
10 repos
Techniques and tools for combining multiple large language models into single unified models, improving performance through strategic merging of model weights and parameters. This cluster covers both foundational merging algorithms and practical tools for practitioners, including specialized approaches for adapters, LoRA weights, and multi-model fusion. Repositories range from core research implementations like mergekit and DAM to downstream applications that leverage merged models for improved inference efficiency and capability combination.
Python · 10
llama ·14,714
llm ·14,714
model-merging ·14,714
