ronantakizawa/lorashare

Python Library for running SHARE (Compress multiple LoRA adapters into a shared subspace)

Python11 stars1 forks

Topics

#lora#low-rank-adaptation#low-rank-approximation#llm-research

What it does

lorashare is a Python library that enables the compression of multiple LoRA adapters into a shared subspace, significantly reducing memory usage. This allows for efficient storage and management of task-specific models without compromising performance.

Star history

Not enough history yet — 1 day(s) recorded. The daily snapshot builds this up.

Tracking

Last trending: 2026-02-21

Creator kit

Hook

Unlock the power of memory-efficient model training with lorashare! Compress multiple LoRA adapters into one with ease.

Content angles

Create a tutorial on how to set up lorashare for model compression.
Discuss the benefits of using shared subspaces in machine learning models.
Share a case study showcasing memory savings achieved using lorashare.

Who should care

AI researchers and developers looking to optimize model storage and performance.