LibreDAMOYOLOt

DAMO-YOLO-T (TinyNAS_res backbone, GiraffeNeckV2, ZeroHead with GFL/DFL), 8.9M params, 640x640 input, post-distill release weights, repackaged for LibreYOLO.

COCO val 2017 mAP: 0.4199 (matches upstream's reported 42.0).

Source

Derived from tinyvision/DAMO-YOLO at commit 319572e (release v0.3.1, post-distill).

The original damoyolo_tinynasL20_T_420.pth checkpoint was hosted on Alibaba's Aliyun bucket which is now offline. This weight was recovered from the Internet Archive's Wayback Machine, snapshot 20250307203525.

Copyright (c) 2021-2022 Alibaba Group Holding Limited. Licensed under the Apache License, Version 2.0.

Modifications

File rename only. damoyolo_tinynasL20_T_420.pth is bit-identical to LibreDAMOYOLOt.pt. LibreYOLO's nn.Module class hierarchy mirrors upstream's exactly so model.load_state_dict(...) is strict.

Verified by running COCO val 2017 with both pipelines: same mAP 0.4199 to four decimals; raw cls_scores and boxes tensors match upstream's output to 0.0 absolute difference on identical input.

See libreyolo/models/damoyolo/ in the LibreYOLO source repository for the architecture port.

License

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

Downloads last month

-

Downloads are not tracked for this model. How to track
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Collection including LibreYOLO/LibreDAMOYOLOt