Downloads · 30 days
16
57% of all-time downloads
LibreYOLO/LibreMoondream3
LibreMoondream3 is a object detection model from LibreYOLO. Use it when you need objects located in an image. It is set up for libreyolo. The card lists the license as other.
Moondream 3 Preview weights mirrored for LibreYOLO's LibreVLM tier.
Downloads · 30 days
16
57% of all-time downloads
All-time downloads
28
Public
Repo size
18.5 GB
Likes
0
Public
Click a slice to open those files.
.safetensors18.5 GB · 100%
From the Hugging Face model README
Moondream 3 Preview weights mirrored for LibreYOLO's LibreVLM tier.
Custom-license weights (BSL 1.1)
These weights are not covered by LibreYOLO's MIT license. They are Moondream 3 (Preview) from M87 Labs, licensed under the Business Source License 1.1 with an Additional Use Grant. Redistribution is allowed. Production use is allowed except offering the model to third parties as a paid hosted or embedded service that competes with M87 Labs's paid versions. That restriction binds you, the downloader. The LibreYOLO code stays MIT. See
LICENSE.
Mirrored from moondream/moondream3-preview
at commit 5112966d1a723413b1c9a1e8bea272b72e647b35.
Copyright (c) M87 Labs, Inc.. Licensed under BSL 1.1.
No learned parameters were changed. This repository preserves the Hugging Face snapshot files needed by LibreYOLO's VLM wrapper, ships the upstream license verbatim, and replaces the model card with LibreYOLO-specific loading notes.
from libreyolo import LibreVLM
model = LibreVLM("moondream-3")
model.set_classes(["person", "helmet"])
results = model.predict("image.jpg")
Official runtime notes ask for about 24 GB of GPU memory.
Business Source License 1.1 with Additional Use Grant. See LICENSE
and NOTICE. Change License is Apache-2.0 two years after first
public release of this version.