| --- |
| license: apache-2.0 |
| library_name: acaua |
| pipeline_tag: object-detection |
| tags: |
| - object-detection |
| - vision |
| - acaua |
| - native-pytorch-port |
| - rtmdet |
| datasets: |
| - coco |
| --- |
| |
| # RTMDet-s — acaua mirror (pure-PyTorch port) |
|
|
| Pure-PyTorch port of RTMDet-s (8.9M params, COCO box AP 44.6) hosted under `CondadosAI/` for use with the [acaua](https://github.com/CondadosAI/acaua) computer vision library. |
|
|
| The architecture has been re-implemented in pure PyTorch under [`acaua.adapters.rtmdet`](https://github.com/CondadosAI/acaua/tree/main/src/acaua/adapters/rtmdet) — no `mmcv`, no `mmengine`, no `mmdet`, no `trust_remote_code`. The weights in this mirror are converted from the upstream mmdet `.pth` checkpoint to safetensors with the acaua adapter's state_dict key naming. They are NOT drop-in compatible with mmdet — they're designed to load cleanly into our `nn.Module` tree. |
| |
| ## Provenance |
| |
| | | | |
| |---|---| |
| | Upstream code | [`open-mmlab/mmdetection`](https://github.com/open-mmlab/mmdetection) @ [`cfd5d3a985`](https://github.com/open-mmlab/mmdetection/tree/cfd5d3a985b0249de009b67d04f37263e11cdf3d) (Apache-2.0) | |
| | Upstream weights URL | `https://download.openmmlab.com/mmdetection/v3.0/rtmdet/rtmdet_s_8xb32-300e_coco/rtmdet_s_8xb32-300e_coco_20220905_161602-387a891e.pth` | |
| | Upstream weights SHA256 | `387a891e157cf0ab57d76b3ffc17bf77247089d672532427930b3140f9e789d6` | |
| | Conversion script | [`scripts/convert_rtmdet.py`](https://github.com/CondadosAI/acaua/blob/main/scripts/convert_rtmdet.py) | |
| | Paper | Lyu et al., *"RTMDet: An Empirical Study of Designing Real-Time Object Detectors"*, arXiv:[2212.07784](https://arxiv.org/abs/2212.07784) | |
| | Mirrored on | 2026-04-20 | |
| | Mirrored by | [CondadosAI/acaua](https://github.com/CondadosAI/acaua) | |
| |
| ## Usage |
| |
| ```python |
| import acaua |
| |
| model = acaua.Model.from_pretrained("CondadosAI/rtmdet_s_coco") |
| results = model.predict("image.jpg") |
| print(results.boxes, results.scores, results.labels) |
| ``` |
| |
| ## License and attribution |
| |
| Redistributed under Apache-2.0, consistent with the upstream code (`open-mmlab/mmdetection`) and the weights released on `download.openmmlab.com`. The acaua adapter is itself a derivative work of the upstream PyTorch implementation — see [`NOTICE`](./NOTICE) for the required attribution chain (code AND weights). |
| |
| ## Citation |
| |
| ```bibtex |
| @misc{lyu2022rtmdet, |
| title={RTMDet: An Empirical Study of Designing Real-Time Object Detectors}, |
| author={Chengqi Lyu and Wenwei Zhang and Haian Huang and Yue Zhou and Yudong Wang and Yanyi Liu and Shilong Zhang and Kai Chen}, |
| year={2022}, |
| eprint={2212.07784}, |
| archivePrefix={arXiv}, |
| primaryClass={cs.CV} |
| } |
| ``` |
| |
