ChaoningZhang/MobileSAM

Health v2 · maintenance only

Flags: no_releases

Adoption not part of the score

5858 stars · 586 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-29, confidence not recorded

MobileSAM is the official implementation of a lightweight version of Meta's Segment Anything Model (SAM), replacing the heavyweight image encoder with a tiny one for fast segmentation on CPUs and mobile devices. It also includes MobileSAMv2 with object-aware prompt sampling for faster segment-everything mode, and supports ONNX export for deployment.

Use cases

• run segment anything model on cpu or mobile devices
• segment objects in images quickly with a lightweight model
• export a segmentation model to onnx for browser or mobile deployment
• add fast auto-labeling segmentation to annotation tools
• integrate sam-based mask generation into image editing or inpainting pipelines

When to choose

• you need SAM-style promptable segmentation but with low latency or limited compute
• you want to run segmentation on CPU, mobile, or in the browser
• you need a drop-in lightweight replacement for the original SAM image encoder

When to avoid

• you need the highest possible segmentation accuracy and can afford the full SAM or larger variants
• you need training/fine-tuning infrastructure rather than inference
• your project depends on features only in newer official SAM releases

Facets

library · maturity stable

machine-learning computer-vision image-processing llm-inference computer-vision image-processing machine-learning artificial-intelligence python cross-platform wasm segment-anything sam lightweight-model onnx-export image-segmentation mobile-inference gpu

Member repositories

For agents

markdown · JSON · MCP: product_card(name="ChaoningZhang/MobileSAM")