Skip to content
ModelListedExternal
Download

LibreDEIMs

DEIM-S (HGNetv2 backbone) repackaged for the LibreYOLO framework.

Sourcehuggingface.co/LibreYOLO/LibreDEIMs/tree/a2aea0c54323041f96dd3552d2b81398729e24fc
KindOtherVersionRevision a2aea0cCgrade

Model source

Kind
Other
File size
39.9–40.1 MB
Version
vRevision a2aea0c
Updated
Oct 14, 2026
Downloads
0
Source
Hugging Face

Source description

LibreDEIMs

DEIM-S (HGNetv2 backbone) repackaged for the LibreYOLO framework.

Source

Detector: Intellindust-AI-Lab/DEIM. Licensed under the Apache License, Version 2.0. Weights derived from upstream deim_hgnetv2_s_coco.pth (DEIM-D-FINE-S, 10M params, 49.0 AP).

DEIM reuses the D-FINE architecture (Peterande/D-FINE) and replaces the classification objective with a Matchability-Aware Loss.

Backbone: HGNetv2 (PP-HGNetV2) from PaddlePaddle/PaddleClas. Copyright (c) 2020 PaddlePaddle Authors. All Rights Reserved. Licensed under the Apache License, Version 2.0.

Modifications

Metadata wrap and file rename only. Learned parameters are unchanged. Conversion script: weights/convert_deim_weights.py.

License

Apache License 2.0. See the LICENSE and NOTICE files in this repository.

Sources

1 source
Hugging FacePrimaryVersion a2aea0c54323041f96dd3552d2b81398729e24fc
huggingface.co/LibreYOLO/LibreDEIMs/tree/a2aea0c54323041f96dd3552d2b81398729e24fc

Verified Aug 26

Model artifacts

2 artifacts

LibreDEIMs.onnx

onnx · 40.1 MB · SHA-256 fb7fb4d4cfa5…4a28 · Hugging Face

Download

LibreDEIMs.pt

pt · 39.9 MB · SHA-256 7c0aa665ead0…8311 · Hugging Face

Download

Source excerpts

2 excerpts
ReadmeAug 26huggingface.co/LibreYOLO/LibreDEIMs

--- license: apache-2.0 library_name: libreyolo tags: - object-detection - deim - d-fine - detr --- # LibreDEIMs DEIM-S (HGNetv2 backbone) repackaged for the [LibreYOLO](https://github.com/LibreYOLO/libreyolo) framework. ## Source Detector: [Intellindust-AI-Lab/DEIM](https://github.com/Intellindust-AI-Lab/DEIM). Licensed under the Apache License, Version 2.0. Weights derived from upstream `deim_hgnetv2_s_coco.pth` (DEIM-D-FINE-S, 10M params, 49.0 AP). DEIM reuses the D-FINE architecture ([Peterande/D-FINE](https://github.com/Peterande/D-FINE)) and replaces the classification objective with a Matchability-Aware Loss. Backbone: HGNetv2 (PP-HGNetV2) from [PaddlePaddle/PaddleClas](https://github.com/PaddlePaddle/PaddleClas). Copyright (c) 2020 PaddlePaddle Authors. All Rights Reserved. Licensed under the Apache License, Version 2.0. ## Modifications Metadata wrap and file rename only. Learned parameters are unchanged. Conversion script: [`weights/convert_deim_weights.py`](https://github.com/LibreYOLO/libreyolo/blob/main/weights/convert_deim_weights.py). ## License Apache License 2.0. See the [`LICENSE`](./LICENSE) and [`NOTICE`](./NOTICE) files in this repository.

NoteAug 26huggingface.co/LibreYOLO/LibreDEIMs

Source context: 0 downloads · 0 likes · Pipeline object-detection · Library libreyolo · Repo LibreYOLO/LibreDEIMs