rtdetr 0.1.0__tar.gz

rtdetr-0.1.0/.gitignore

@@ -0,0 +1,16 @@
+__pycache__/
+*.py[cod]
+*.egg-info/
+build/
+dist/
+.venv/
+venv/
+.pytest_cache/
+.ruff_cache/
+runs/
+*.onnx
+*.xml
+*.bin
+*.pt
+!tests/**/*.xml
+.DS_Store

rtdetr-0.1.0/LICENSE

@@ -0,0 +1,202 @@
+
+                                 Apache License
+                           Version 2.0, January 2004
+                        http://www.apache.org/licenses/
+
+   TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
+
+   1. Definitions.
+
+      "License" shall mean the terms and conditions for use, reproduction,
+      and distribution as defined by Sections 1 through 9 of this document.
+
+      "Licensor" shall mean the copyright owner or entity authorized by
+      the copyright owner that is granting the License.
+
+      "Legal Entity" shall mean the union of the acting entity and all
+      other entities that control, are controlled by, or are under common
+      control with that entity. For the purposes of this definition,
+      "control" means (i) the power, direct or indirect, to cause the
+      direction or management of such entity, whether by contract or
+      otherwise, or (ii) ownership of fifty percent (50%) or more of the
+      outstanding shares, or (iii) beneficial ownership of such entity.
+
+      "You" (or "Your") shall mean an individual or Legal Entity
+      exercising permissions granted by this License.
+
+      "Source" form shall mean the preferred form for making modifications,
+      including but not limited to software source code, documentation
+      source, and configuration files.
+
+      "Object" form shall mean any form resulting from mechanical
+      transformation or translation of a Source form, including but
+      not limited to compiled object code, generated documentation,
+      and conversions to other media types.
+
+      "Work" shall mean the work of authorship, whether in Source or
+      Object form, made available under the License, as indicated by a
+      copyright notice that is included in or attached to the work
+      (an example is provided in the Appendix below).
+
+      "Derivative Works" shall mean any work, whether in Source or Object
+      form, that is based on (or derived from) the Work and for which the
+      editorial revisions, annotations, elaborations, or other modifications
+      represent, as a whole, an original work of authorship. For the purposes
+      of this License, Derivative Works shall not include works that remain
+      separable from, or merely link (or bind by name) to the interfaces of,
+      the Work and Derivative Works thereof.
+
+      "Contribution" shall mean any work of authorship, including
+      the original version of the Work and any modifications or additions
+      to that Work or Derivative Works thereof, that is intentionally
+      submitted to Licensor for inclusion in the Work by the copyright owner
+      or by an individual or Legal Entity authorized to submit on behalf of
+      the copyright owner. For the purposes of this definition, "submitted"
+      means any form of electronic, verbal, or written communication sent
+      to the Licensor or its representatives, including but not limited to
+      communication on electronic mailing lists, source code control systems,
+      and issue tracking systems that are managed by, or on behalf of, the
+      Licensor for the purpose of discussing and improving the Work, but
+      excluding communication that is conspicuously marked or otherwise
+      designated in writing by the copyright owner as "Not a Contribution."
+
+      "Contributor" shall mean Licensor and any individual or Legal Entity
+      on behalf of whom a Contribution has been received by Licensor and
+      subsequently incorporated within the Work.
+
+   2. Grant of Copyright License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      copyright license to reproduce, prepare Derivative Works of,
+      publicly display, publicly perform, sublicense, and distribute the
+      Work and such Derivative Works in Source or Object form.
+
+   3. Grant of Patent License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      (except as stated in this section) patent license to make, have made,
+      use, offer to sell, sell, import, and otherwise transfer the Work,
+      where such license applies only to those patent claims licensable
+      by such Contributor that are necessarily infringed by their
+      Contribution(s) alone or by combination of their Contribution(s)
+      with the Work to which such Contribution(s) was submitted. If You
+      institute patent litigation against any entity (including a
+      cross-claim or counterclaim in a lawsuit) alleging that the Work
+      or a Contribution incorporated within the Work constitutes direct
+      or contributory patent infringement, then any patent licenses
+      granted to You under this License for that Work shall terminate
+      as of the date such litigation is filed.
+
+   4. Redistribution. You may reproduce and distribute copies of the
+      Work or Derivative Works thereof in any medium, with or without
+      modifications, and in Source or Object form, provided that You
+      meet the following conditions:
+
+      (a) You must give any other recipients of the Work or
+          Derivative Works a copy of this License; and
+
+      (b) You must cause any modified files to carry prominent notices
+          stating that You changed the files; and
+
+      (c) You must retain, in the Source form of any Derivative Works
+          that You distribute, all copyright, patent, trademark, and
+          attribution notices from the Source form of the Work,
+          excluding those notices that do not pertain to any part of
+          the Derivative Works; and
+
+      (d) If the Work includes a "NOTICE" text file as part of its
+          distribution, then any Derivative Works that You distribute must
+          include a readable copy of the attribution notices contained
+          within such NOTICE file, excluding those notices that do not
+          pertain to any part of the Derivative Works, in at least one
+          of the following places: within a NOTICE text file distributed
+          as part of the Derivative Works; within the Source form or
+          documentation, if provided along with the Derivative Works; or,
+          within a display generated by the Derivative Works, if and
+          wherever such third-party notices normally appear. The contents
+          of the NOTICE file are for informational purposes only and
+          do not modify the License. You may add Your own attribution
+          notices within Derivative Works that You distribute, alongside
+          or as an addendum to the NOTICE text from the Work, provided
+          that such additional attribution notices cannot be construed
+          as modifying the License.
+
+      You may add Your own copyright statement to Your modifications and
+      may provide additional or different license terms and conditions
+      for use, reproduction, or distribution of Your modifications, or
+      for any such Derivative Works as a whole, provided Your use,
+      reproduction, and distribution of the Work otherwise complies with
+      the conditions stated in this License.
+
+   5. Submission of Contributions. Unless You explicitly state otherwise,
+      any Contribution intentionally submitted for inclusion in the Work
+      by You to the Licensor shall be under the terms and conditions of
+      this License, without any additional terms or conditions.
+      Notwithstanding the above, nothing herein shall supersede or modify
+      the terms of any separate license agreement you may have executed
+      with Licensor regarding such Contributions.
+
+   6. Trademarks. This License does not grant permission to use the trade
+      names, trademarks, service marks, or product names of the Licensor,
+      except as required for reasonable and customary use in describing the
+      origin of the Work and reproducing the content of the NOTICE file.
+
+   7. Disclaimer of Warranty. Unless required by applicable law or
+      agreed to in writing, Licensor provides the Work (and each
+      Contributor provides its Contributions) on an "AS IS" BASIS,
+      WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
+      implied, including, without limitation, any warranties or conditions
+      of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
+      PARTICULAR PURPOSE. You are solely responsible for determining the
+      appropriateness of using or redistributing the Work and assume any
+      risks associated with Your exercise of permissions under this License.
+
+   8. Limitation of Liability. In no event and under no legal theory,
+      whether in tort (including negligence), contract, or otherwise,
+      unless required by applicable law (such as deliberate and grossly
+      negligent acts) or agreed to in writing, shall any Contributor be
+      liable to You for damages, including any direct, indirect, special,
+      incidental, or consequential damages of any character arising as a
+      result of this License or out of the use or inability to use the
+      Work (including but not limited to damages for loss of goodwill,
+      work stoppage, computer failure or malfunction, or any and all
+      other commercial damages or losses), even if such Contributor
+      has been advised of the possibility of such damages.
+
+   9. Accepting Warranty or Additional Liability. While redistributing
+      the Work or Derivative Works thereof, You may choose to offer,
+      and charge a fee for, acceptance of support, warranty, indemnity,
+      or other liability obligations and/or rights consistent with this
+      License. However, in accepting such obligations, You may act only
+      on Your own behalf and on Your sole responsibility, not on behalf
+      of any other Contributor, and only if You agree to indemnify,
+      defend, and hold each Contributor harmless for any liability
+      incurred by, or claims asserted against, such Contributor by reason
+      of your accepting any such warranty or additional liability.
+
+   END OF TERMS AND CONDITIONS
+
+   APPENDIX: How to apply the Apache License to your work.
+
+      To apply the Apache License to your work, attach the following
+      boilerplate notice, with the fields enclosed by brackets "[]"
+      replaced with your own identifying information. (Don't include
+      the brackets!)  The text should be enclosed in the appropriate
+      comment syntax for the file format. We also recommend that a
+      file or class name and description of purpose be included on the
+      same "printed page" as the copyright notice for easier
+      identification within third-party archives.
+
+   Copyright 2026 rtdetr contributors
+
+   Licensed under the Apache License, Version 2.0 (the "License");
+   you may not use this file except in compliance with the License.
+   You may obtain a copy of the License at
+
+       http://www.apache.org/licenses/LICENSE-2.0
+
+   Unless required by applicable law or agreed to in writing, software
+   distributed under the License is distributed on an "AS IS" BASIS,
+   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+   See the License for the specific language governing permissions and
+   limitations under the License.

rtdetr-0.1.0/PKG-INFO

@@ -0,0 +1,277 @@
+Metadata-Version: 2.5
+Name: rtdetr
+Version: 0.1.0
+Summary: RT-DETR object detection with an Ultralytics-style API — 100% Apache-2.0
+Project-URL: Homepage, https://github.com/leeyunjai82/rtdetr
+Project-URL: Source, https://github.com/leeyunjai82/rtdetr
+Project-URL: Issues, https://github.com/leeyunjai82/rtdetr/issues
+Author-email: leeyunjai <leeyunjai1982@gmail.com>
+License-Expression: Apache-2.0
+License-File: LICENSE
+Keywords: apache-2.0,detr,object-detection,openvino,rt-detr,rtdetr
+Classifier: Development Status :: 4 - Beta
+Classifier: Intended Audience :: Developers
+Classifier: Intended Audience :: Science/Research
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
+Requires-Python: >=3.10
+Requires-Dist: numpy>=1.23
+Requires-Dist: opencv-python>=4.6
+Requires-Dist: openvino>=2024.0
+Requires-Dist: pyyaml>=5.4
+Provides-Extra: dev
+Requires-Dist: pillow>=9.0; extra == 'dev'
+Requires-Dist: pytest>=7.0; extra == 'dev'
+Requires-Dist: ruff>=0.5; extra == 'dev'
+Provides-Extra: train
+Requires-Dist: onnx>=1.14; extra == 'train'
+Requires-Dist: scipy>=1.9; extra == 'train'
+Requires-Dist: torch>=2.1; extra == 'train'
+Requires-Dist: torchvision>=0.16; extra == 'train'
+Description-Content-Type: text/markdown
+
+# rtdetr
+
+**Ultralytics-style API, 100% Apache-2.0.** RT-DETR object detection you can ship
+inside a product: train with PyTorch, run with OpenVINO, and keep the workflow
+your team already knows — `model.train(...)`, `model.predict(...)`,
+`results[0].boxes.xyxy`.
+
+Every line of the network, loss, trainer, validator and exporter is an original
+implementation. No Ultralytics code, no AGPL weights, no license surprises.
+
+```bash
+pip install rtdetr              # inference (numpy, opencv, openvino, pyyaml)
+pip install "rtdetr[train]"     # + training (torch, torchvision, scipy, onnx)
+```
+
+```python
+from rtdetr import RTDETR
+
+model = RTDETR("rtdetr-r18")          # weights from the mirror (see "Pretrained weights")
+results = model("bus.jpg", conf=0.5)  # list[Results]
+
+r = results[0]
+r.boxes.xyxy, r.boxes.conf, r.boxes.cls    # plain numpy
+r.names[int(r.boxes.cls[0])]               # "person"
+r.plot(); r.save(); r.show()
+```
+
+```python
+model = RTDETR("rtdetr-r18")
+model.train(data="data.yaml", epochs=100, imgsz=640, batch=8, device=0)
+metrics = model.val(data="data.yaml")       # metrics.box.map50 / .box.map
+model.export(format="openvino", half=True)  # IR + labels.txt, ready to deploy
+```
+
+---
+
+## Why this exists
+
+Ultralytics YOLO is excellent and its API is muscle memory for a lot of people —
+but the code and weights are AGPL-3.0, which rules them out of most shipped
+products. `rtdetr` keeps the ergonomics and drops the license problem.
+
+## Predicting
+
+Any source a YOLO user would expect:
+
+| Source | Example |
+| --- | --- |
+| image | `model("bus.jpg")` |
+| glob | `model("frames/*.jpg")` |
+| folder | `model("dataset/images")` |
+| list file | `model("images.txt")` |
+| URL | `model("https://example.com/bus.jpg")` |
+| video | `model("clip.mp4")` |
+| stream | `model("rtsp://camera/live")` |
+| webcam | `model(0)` |
+| array | `model(numpy_bgr)` / `model(pil_image)` |
+| list | `model(["a.jpg", "b.jpg"])` |
+
+```python
+for r in model.predict("clip.mp4", conf=0.4, stream=True):   # generator, O(1) memory
+    print(r.boxes.xyxyn)
+
+model.predict("bus.jpg", save=True)     # writes runs/detect/predict/bus.jpg
+model.track("clip.mp4")                 # IoU tracker -> r.boxes.id
+```
+
+The log line reads the way you expect:
+
+```
+image 1/1 bus.jpg: 640x640 4 persons, 1 bus, 12.3ms
+```
+
+### Results
+
+| Attribute | What you get |
+| --- | --- |
+| `r.boxes.xyxy` | `(N, 4)` pixel corners |
+| `r.boxes.xywh` | `(N, 4)` pixel centre + size |
+| `r.boxes.xyxyn` / `r.boxes.xywhn` | the same, normalized 0..1 |
+| `r.boxes.conf` / `r.boxes.cls` | `(N,)` scores and class indices |
+| `r.boxes.id` | track ids after `model.track(...)`, else `None` |
+| `r.names` | `{0: "person", ...}` |
+| `r.plot()` | annotated BGR ndarray |
+| `r.save()` / `r.show()` | write / display it |
+| `r.summary()` | detections as JSON-ready dicts |
+| `r.speed` | `{"preprocess": ms, "inference": ms, "postprocess": ms}` |
+
+## Training
+
+YOLO-format labels — the same `data.yaml` and `images/` + `labels/*.txt` layout:
+
+```yaml
+path: /data/cans
+train: images/train
+val: images/val
+names:
+  0: can
+  1: bottle
+```
+
+```python
+model = RTDETR("rtdetr-r18")
+best = model.train(
+    data="data.yaml", epochs=100, imgsz=640, batch=8,
+    device=0, workers=4, project="runs", name="train",
+    resume=False, patience=50, lr0=1e-4, seed=0,
+)
+# runs/train/weights/best.pt  (last.pt too — resume=True picks the run back up)
+```
+
+AdamW with a 10× lower LR on the backbone, linear warmup into cosine decay, AMP
+on CUDA, mAP50-95 after every epoch, early stop on `patience`.
+
+## Exporting and deploying
+
+```python
+model = RTDETR("runs/train/weights/best.pt")
+xml = model.export(format="openvino", half=True, imgsz=640)
+```
+
+Writes `best.xml` + `best.bin`, plus **`labels.txt`** (one class name per line)
+next to the IR — that is how downstream runtimes discover class names.
+
+The exported graph emits **probabilities, not logits**. Anything that decodes it
+must not apply a sigmoid a second time; this package checks the score range and
+skips it (`rtdetr/predictor.py`).
+
+## CLI
+
+Same `key=value` shape as the `yolo` command:
+
+```bash
+rtdetr predict model=rtdetr-r18 source=bus.jpg conf=0.5
+rtdetr train   model=rtdetr-r18 data=data.yaml epochs=100
+rtdetr val     model=best.pt data=data.yaml
+rtdetr export  model=best.pt format=openvino half=true
+rtdetr track   model=best.pt source=clip.mp4
+```
+
+## Pretrained weights
+
+> **Status: the mirror is not populated yet.** Until COCO weights are published
+> there, `RTDETR("rtdetr-r18")` raises a `ModelNotFoundError` that names the ways
+> forward — it never silently falls back to an untrained network. Everything else
+> (your own `.pt`/`.xml`, training, val, export) works today.
+
+`RTDETR("rtdetr-r18")` downloads the IR for that name on first use and caches it
+in `~/.rtdetr/` (`$RTDETR_HOME` to move it, `$RTDETR_ASSETS_URL` to point at an
+internal mirror — handy for air-gapped sites). Known names: `rtdetr-r18`,
+`rtdetr-r34`, `rtdetr-r50`. Each mirror entry is `<name>/<name>.xml`, `.bin`,
+`.pt` and `labels.txt`.
+
+Publishing an entry takes three steps, and only the first is instant:
+
+1. `tools/convert_official.py` — move what maps from the original release.
+2. Fine-tune the result on COCO. The converted decoder cross-attention and CCFF
+   blocks start fresh (see below), so this step is what actually earns the
+   "COCO pretrained" label.
+3. Upload `.pt`, `.xml`, `.bin` and `labels.txt` to the mirror path.
+
+`tools/convert_official.py` converts the original
+[lyuwenyu/RT-DETR](https://github.com/lyuwenyu/RT-DETR) Apache-2.0 COCO weights
+into this package's layout:
+
+```bash
+python tools/convert_official.py --weights rtdetr_r18vd_6x_coco.pth \
+    --variant r18 --out rtdetr-r18.pt --report 10
+```
+
+It prints exactly what transferred. Note the honest part: this decoder uses plain
+multi-head cross-attention rather than deformable attention (that is what keeps
+ONNX/OpenVINO export dependency-free), and the CCFF fusion blocks are narrower,
+so those tensors have no counterpart upstream. **A converted checkpoint is a warm
+start, not a finished COCO model — fine-tune it before publishing.**
+
+## Design notes
+
+* **Backbone** ResNet-18/34/50, torchvision-compatible module names, so ImageNet
+  weights load when torchvision is around.
+* **Encoder** AIFI transformer on C5 + FPN/PAN cross-scale fusion.
+* **Decoder** two-stage: dense heads pick the top-K encoder tokens as queries, six
+  layers refine them. No deformable attention, on purpose — see above.
+* **Loss** Hungarian matching, varifocal + L1 + GIoU, auxiliary and encoder heads.
+* **Validation** COCO-style mAP50 / mAP50-95, 101-point interpolation, no
+  pycocotools dependency.
+* **Inference** OpenVINO, plain resize (no letterbox) to match training.
+
+## Development
+
+```bash
+pip install -e ".[train,dev]"
+pytest -q          # the parity checklist lives in tests/
+ruff check .
+```
+
+## Releasing
+
+Push a `v*` tag and GitHub Actions builds and uploads to PyPI:
+
+```bash
+git tag v0.1.0
+git push origin v0.1.0
+```
+
+`.github/workflows/publish.yml` uses **PyPI trusted publishing** — no API token
+lives in the repo. One-time setup on PyPI (Publishing → add a pending publisher):
+
+| Field | Value |
+| --- | --- |
+| PyPI project | `rtdetr` |
+| Owner | `leeyunjai82` |
+| Repository | `rtdetr` |
+| Workflow | `publish.yml` |
+| Environment | `pypi` |
+
+The job refuses to publish when the tag and `pyproject.toml` version disagree, so
+bump the version in the same commit you tag.
+
+## 한국어 빠른 시작
+
+Ultralytics YOLO와 사용법이 같지만 라이선스는 Apache-2.0입니다. 제품에 넣어도
+문제없습니다.
+
+```python
+from rtdetr import RTDETR
+
+model = RTDETR("rtdetr-r18")               # 미러에서 가중치 다운로드 (아래 "Pretrained weights" 참고)
+results = model("bus.jpg", conf=0.5)       # list[Results]
+results[0].boxes.xyxy                      # numpy 배열
+results[0].save()                          # 결과 이미지 저장
+
+model.train(data="data.yaml", epochs=100)  # YOLO 형식 라벨 그대로
+model.val(data="data.yaml").box.map50      # mAP50
+model.export(format="openvino", half=True) # IR + labels.txt
+```
+
+가중치 캐시는 `~/.rtdetr/`, 사내 미러는 `RTDETR_ASSETS_URL` 환경변수로 지정합니다.
+
+## License
+
+Apache-2.0. See [LICENSE](LICENSE).