rfdetr-plus 0.0.0.dev0__tar.gz → 0.0.1__tar.gz

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
@@ -0,0 +1,162 @@
1
+ Metadata-Version: 2.4
2
+ Name: rfdetr-plus
3
+ Version: 0.0.1
4
+ Summary: RF-DETR+ Extension Package
5
+ Author-email: "Roboflow, Inc" <peter@roboflow.com>
6
+ License-Expression: LicenseRef-PML-1.0
7
+ Project-URL: Homepage, https://github.com/roboflow/rf-detr-plus
8
+ Keywords: machine-learning,deep-learning,vision,ML,DL,AI,DETR,RF-DETR,Roboflow
9
+ Classifier: Development Status :: 3 - Alpha
10
+ Classifier: Intended Audience :: Developers
11
+ Classifier: Intended Audience :: Education
12
+ Classifier: Intended Audience :: Science/Research
13
+ Classifier: Programming Language :: Python :: 3
14
+ Classifier: Programming Language :: Python :: 3.10
15
+ Classifier: Programming Language :: Python :: 3.11
16
+ Classifier: Programming Language :: Python :: 3.12
17
+ Classifier: Programming Language :: Python :: 3.13
18
+ Classifier: Programming Language :: Python :: 3 :: Only
19
+ Classifier: Topic :: Software Development
20
+ Classifier: Topic :: Scientific/Engineering
21
+ Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
22
+ Classifier: Typing :: Typed
23
+ Classifier: Operating System :: POSIX
24
+ Classifier: Operating System :: Unix
25
+ Classifier: Operating System :: MacOS
26
+ Requires-Python: >=3.10
27
+ Description-Content-Type: text/markdown
28
+ License-File: LICENSE
29
+ Requires-Dist: rfdetr<2,>=1.4.1
30
+ Dynamic: license-file
31
+
32
+ # RF-DETR+: Large-Scale Detection Models for RF-DETR
33
+
34
+ [![version](https://badge.fury.io/py/rfdetr-plus.svg)](https://badge.fury.io/py/rfdetr-plus)
35
+ [![downloads](https://img.shields.io/pypi/dm/rfdetr-plus)](https://pypistats.org/packages/rfdetr-plus)
36
+ [![python-version](https://img.shields.io/pypi/pyversions/rfdetr-plus)](https://badge.fury.io/py/rfdetr-plus)
37
+ [![license](https://img.shields.io/badge/license-PML--1.0-blue)](https://github.com/roboflow/rf-detr-plus/blob/main/LICENSE)
38
+
39
+ [![discord](https://img.shields.io/discord/1159501506232451173?logo=discord&label=discord&labelColor=fff&color=5865f2&link=https%3A%2F%2Fdiscord.gg%2FGbfgXGJ8Bk)](https://discord.gg/GbfgXGJ8Bk)
40
+
41
+ RF-DETR+ is the extension package for [RF-DETR](https://github.com/roboflow/rf-detr) that provides the **XLarge** and **2XLarge** detection models. These are the largest and most accurate models in the RF-DETR family, pushing state-of-the-art accuracy on [Microsoft COCO](https://cocodataset.org/#home) and [RF100-VL](https://github.com/roboflow/rf100-vl) while retaining real-time inference speeds.
42
+
43
+ RF-DETR+ models use a DINOv2 vision transformer backbone at higher resolutions and larger feature dimensions than the core RF-DETR lineup, unlocking top-tier detection accuracy for applications where precision matters most.
44
+
45
+ ## Install
46
+
47
+ Install RF-DETR+ in a [**Python>=3.10**](https://www.python.org/) environment with `pip`. This will also install [`rfdetr`](https://github.com/roboflow/rf-detr) as a dependency.
48
+
49
+ ```bash
50
+ pip install rfdetr-plus
51
+ ```
52
+
53
+ <details>
54
+ <summary>Install from source</summary>
55
+
56
+ <br>
57
+
58
+ ```bash
59
+ pip install git+https://github.com/roboflow/rf-detr-plus.git
60
+ ```
61
+
62
+ </details>
63
+
64
+ ## Benchmarks
65
+
66
+ RF-DETR+ XLarge and 2XLarge sit at the top of the RF-DETR accuracy/latency curve, delivering the highest COCO AP scores in the family. All latency numbers were measured on an NVIDIA T4 using TensorRT, FP16, and batch size 1.
67
+
68
+ | Size | Class | COCO AP<sub>50</sub> | COCO AP<sub>50:95</sub> | RF100VL AP<sub>50</sub> | RF100VL AP<sub>50:95</sub> | Latency (ms) | Params (M) | Resolution | Package / License |
69
+ | :----: | :-------------: | :------------------: | :---------------------: | :---------------------: | :------------------------: | :----------: | :--------: | :--------: | :----------------------------------------------------------: |
70
+ | N | `RFDETRNano` | 67.6 | 48.4 | 85.0 | 57.7 | 2.3 | 30.5 | 384x384 | [`rfdetr`](https://github.com/roboflow/rf-detr) / Apache 2.0 |
71
+ | S | `RFDETRSmall` | 72.1 | 53.0 | 86.7 | 60.2 | 3.5 | 32.1 | 512x512 | [`rfdetr`](https://github.com/roboflow/rf-detr) / Apache 2.0 |
72
+ | M | `RFDETRMedium` | 73.6 | 54.7 | 87.4 | 61.2 | 4.4 | 33.7 | 576x576 | [`rfdetr`](https://github.com/roboflow/rf-detr) / Apache 2.0 |
73
+ | L | `RFDETRLarge` | 75.1 | 56.5 | 88.2 | 62.2 | 6.8 | 33.9 | 704x704 | [`rfdetr`](https://github.com/roboflow/rf-detr) / Apache 2.0 |
74
+ | ⭐ XL | `RFDETRXLarge` | 77.4 | 58.6 | 88.5 | 62.9 | 11.5 | 126.4 | 700x700 | `rfdetr_plus` / [PML 1.0](LICENSE) |
75
+ | ⭐ 2XL | `RFDETR2XLarge` | 78.5 | 60.1 | 89.0 | 63.2 | 17.2 | 126.9 | 880x880 | `rfdetr_plus` / [PML 1.0](LICENSE) |
76
+
77
+ ## Run Models
78
+
79
+ Install with the `plus` extra to get XL and 2XL models alongside core RF-DETR:
80
+
81
+ ```bash
82
+ pip install rfdetr_plus
83
+ ```
84
+
85
+ RF-DETR+ models require you to accept the Platform Model License before use. Once accepted, usage mirrors the standard RF-DETR API -- you import directly from `rfdetr`:
86
+
87
+ ```python
88
+ import requests
89
+ import supervision as sv
90
+ from PIL import Image
91
+ from rfdetr import RFDETRXLarge
92
+ from rfdetr.util.coco_classes import COCO_CLASSES
93
+
94
+ model = RFDETRXLarge(accept_platform_model_license=True)
95
+
96
+ image = Image.open(requests.get("https://media.roboflow.com/dog.jpg", stream=True).raw)
97
+ detections = model.predict(image, threshold=0.5)
98
+
99
+ labels = [f"{COCO_CLASSES[class_id]}" for class_id in detections.class_id]
100
+
101
+ annotated_image = sv.BoxAnnotator().annotate(image, detections)
102
+ annotated_image = sv.LabelAnnotator().annotate(annotated_image, detections, labels)
103
+ ```
104
+
105
+ ### Train Models
106
+
107
+ RF-DETR+ models support fine-tuning with the same training API as core RF-DETR. You can train on your own dataset or use datasets from [Roboflow Universe](https://universe.roboflow.com/).
108
+
109
+ ```python
110
+ from rfdetr import RFDETRXLarge
111
+
112
+ model = RFDETRXLarge(accept_platform_model_license=True)
113
+ model.train(dataset_dir="path/to/dataset", epochs=50, lr=1e-4)
114
+ ```
115
+
116
+ ## Documentation
117
+
118
+ Visit the [RF-DETR documentation website](https://rfdetr.roboflow.com) to learn more about training, export, deployment, and the full model lineup.
119
+
120
+ ## License
121
+
122
+ RF-DETR+ code and model checkpoints are licensed under the Platform Model License 1.0 (PML-1.0). See [`LICENSE`](LICENSE) for details. These models require a [Roboflow](https://roboflow.com) account to run and fine-tune.
123
+
124
+ The core RF-DETR models (Nano through Large) are available under the Apache License 2.0 in the [`rfdetr`](https://github.com/roboflow/rf-detr) package.
125
+
126
+ ## Acknowledgements
127
+
128
+ Our work is built upon [LW-DETR](https://arxiv.org/pdf/2406.03459), [DINOv2](https://arxiv.org/pdf/2304.07193), and [Deformable DETR](https://arxiv.org/pdf/2010.04159). Thanks to their authors for their excellent work!
129
+
130
+ ## Citation
131
+
132
+ If you find our work helpful for your research, please consider citing the following BibTeX entry.
133
+
134
+ ```bibtex
135
+ @misc{rf-detr,
136
+ title={RF-DETR: Neural Architecture Search for Real-Time Detection Transformers},
137
+ author={Isaac Robinson and Peter Robicheaux and Matvei Popov and Deva Ramanan and Neehar Peri},
138
+ year={2025},
139
+ eprint={2511.09554},
140
+ archivePrefix={arXiv},
141
+ primaryClass={cs.CV},
142
+ url={https://arxiv.org/abs/2511.09554},
143
+ }
144
+ ```
145
+
146
+ ## Contribute
147
+
148
+ We welcome and appreciate all contributions! If you notice any issues or bugs, have questions, or would like to suggest new features, please [open an issue](https://github.com/roboflow/rf-detr-plus/issues/new) or pull request. By sharing your ideas and improvements, you help make RF-DETR better for everyone.
149
+
150
+ <p align="center">
151
+ <a href="https://youtube.com/roboflow"><img src="https://media.roboflow.com/notebooks/template/icons/purple/youtube.png?ik-sdk-version=javascript-1.4.3&updatedAt=1672949634652" width="3%"/></a>
152
+ <img src="https://raw.githubusercontent.com/ultralytics/assets/main/social/logo-transparent.png" width="3%"/>
153
+ <a href="https://roboflow.com"><img src="https://media.roboflow.com/notebooks/template/icons/purple/roboflow-app.png?ik-sdk-version=javascript-1.4.3&updatedAt=1672949746649" width="3%"/></a>
154
+ <img src="https://raw.githubusercontent.com/ultralytics/assets/main/social/logo-transparent.png" width="3%"/>
155
+ <a href="https://www.linkedin.com/company/roboflow-ai/"><img src="https://media.roboflow.com/notebooks/template/icons/purple/linkedin.png?ik-sdk-version=javascript-1.4.3&updatedAt=1672949633691" width="3%"/></a>
156
+ <img src="https://raw.githubusercontent.com/ultralytics/assets/main/social/logo-transparent.png" width="3%"/>
157
+ <a href="https://docs.roboflow.com"><img src="https://media.roboflow.com/notebooks/template/icons/purple/knowledge.png?ik-sdk-version=javascript-1.4.3&updatedAt=1672949634511" width="3%"/></a>
158
+ <img src="https://raw.githubusercontent.com/ultralytics/assets/main/social/logo-transparent.png" width="3%"/>
159
+ <a href="https://discuss.roboflow.com"><img src="https://media.roboflow.com/notebooks/template/icons/purple/forum.png?ik-sdk-version=javascript-1.4.3&updatedAt=1672949633584" width="3%"/></a>
160
+ <img src="https://raw.githubusercontent.com/ultralytics/assets/main/social/logo-transparent.png" width="3%"/>
161
+ <a href="https://blog.roboflow.com"><img src="https://media.roboflow.com/notebooks/template/icons/purple/blog.png?ik-sdk-version=javascript-1.4.3&updatedAt=1672949633605" width="3%"/></a>
162
+ </p>
@@ -0,0 +1,131 @@
1
+ # RF-DETR+: Large-Scale Detection Models for RF-DETR
2
+
3
+ [![version](https://badge.fury.io/py/rfdetr-plus.svg)](https://badge.fury.io/py/rfdetr-plus)
4
+ [![downloads](https://img.shields.io/pypi/dm/rfdetr-plus)](https://pypistats.org/packages/rfdetr-plus)
5
+ [![python-version](https://img.shields.io/pypi/pyversions/rfdetr-plus)](https://badge.fury.io/py/rfdetr-plus)
6
+ [![license](https://img.shields.io/badge/license-PML--1.0-blue)](https://github.com/roboflow/rf-detr-plus/blob/main/LICENSE)
7
+
8
+ [![discord](https://img.shields.io/discord/1159501506232451173?logo=discord&label=discord&labelColor=fff&color=5865f2&link=https%3A%2F%2Fdiscord.gg%2FGbfgXGJ8Bk)](https://discord.gg/GbfgXGJ8Bk)
9
+
10
+ RF-DETR+ is the extension package for [RF-DETR](https://github.com/roboflow/rf-detr) that provides the **XLarge** and **2XLarge** detection models. These are the largest and most accurate models in the RF-DETR family, pushing state-of-the-art accuracy on [Microsoft COCO](https://cocodataset.org/#home) and [RF100-VL](https://github.com/roboflow/rf100-vl) while retaining real-time inference speeds.
11
+
12
+ RF-DETR+ models use a DINOv2 vision transformer backbone at higher resolutions and larger feature dimensions than the core RF-DETR lineup, unlocking top-tier detection accuracy for applications where precision matters most.
13
+
14
+ ## Install
15
+
16
+ Install RF-DETR+ in a [**Python>=3.10**](https://www.python.org/) environment with `pip`. This will also install [`rfdetr`](https://github.com/roboflow/rf-detr) as a dependency.
17
+
18
+ ```bash
19
+ pip install rfdetr-plus
20
+ ```
21
+
22
+ <details>
23
+ <summary>Install from source</summary>
24
+
25
+ <br>
26
+
27
+ ```bash
28
+ pip install git+https://github.com/roboflow/rf-detr-plus.git
29
+ ```
30
+
31
+ </details>
32
+
33
+ ## Benchmarks
34
+
35
+ RF-DETR+ XLarge and 2XLarge sit at the top of the RF-DETR accuracy/latency curve, delivering the highest COCO AP scores in the family. All latency numbers were measured on an NVIDIA T4 using TensorRT, FP16, and batch size 1.
36
+
37
+ | Size | Class | COCO AP<sub>50</sub> | COCO AP<sub>50:95</sub> | RF100VL AP<sub>50</sub> | RF100VL AP<sub>50:95</sub> | Latency (ms) | Params (M) | Resolution | Package / License |
38
+ | :----: | :-------------: | :------------------: | :---------------------: | :---------------------: | :------------------------: | :----------: | :--------: | :--------: | :----------------------------------------------------------: |
39
+ | N | `RFDETRNano` | 67.6 | 48.4 | 85.0 | 57.7 | 2.3 | 30.5 | 384x384 | [`rfdetr`](https://github.com/roboflow/rf-detr) / Apache 2.0 |
40
+ | S | `RFDETRSmall` | 72.1 | 53.0 | 86.7 | 60.2 | 3.5 | 32.1 | 512x512 | [`rfdetr`](https://github.com/roboflow/rf-detr) / Apache 2.0 |
41
+ | M | `RFDETRMedium` | 73.6 | 54.7 | 87.4 | 61.2 | 4.4 | 33.7 | 576x576 | [`rfdetr`](https://github.com/roboflow/rf-detr) / Apache 2.0 |
42
+ | L | `RFDETRLarge` | 75.1 | 56.5 | 88.2 | 62.2 | 6.8 | 33.9 | 704x704 | [`rfdetr`](https://github.com/roboflow/rf-detr) / Apache 2.0 |
43
+ | ⭐ XL | `RFDETRXLarge` | 77.4 | 58.6 | 88.5 | 62.9 | 11.5 | 126.4 | 700x700 | `rfdetr_plus` / [PML 1.0](LICENSE) |
44
+ | ⭐ 2XL | `RFDETR2XLarge` | 78.5 | 60.1 | 89.0 | 63.2 | 17.2 | 126.9 | 880x880 | `rfdetr_plus` / [PML 1.0](LICENSE) |
45
+
46
+ ## Run Models
47
+
48
+ Install with the `plus` extra to get XL and 2XL models alongside core RF-DETR:
49
+
50
+ ```bash
51
+ pip install rfdetr_plus
52
+ ```
53
+
54
+ RF-DETR+ models require you to accept the Platform Model License before use. Once accepted, usage mirrors the standard RF-DETR API -- you import directly from `rfdetr`:
55
+
56
+ ```python
57
+ import requests
58
+ import supervision as sv
59
+ from PIL import Image
60
+ from rfdetr import RFDETRXLarge
61
+ from rfdetr.util.coco_classes import COCO_CLASSES
62
+
63
+ model = RFDETRXLarge(accept_platform_model_license=True)
64
+
65
+ image = Image.open(requests.get("https://media.roboflow.com/dog.jpg", stream=True).raw)
66
+ detections = model.predict(image, threshold=0.5)
67
+
68
+ labels = [f"{COCO_CLASSES[class_id]}" for class_id in detections.class_id]
69
+
70
+ annotated_image = sv.BoxAnnotator().annotate(image, detections)
71
+ annotated_image = sv.LabelAnnotator().annotate(annotated_image, detections, labels)
72
+ ```
73
+
74
+ ### Train Models
75
+
76
+ RF-DETR+ models support fine-tuning with the same training API as core RF-DETR. You can train on your own dataset or use datasets from [Roboflow Universe](https://universe.roboflow.com/).
77
+
78
+ ```python
79
+ from rfdetr import RFDETRXLarge
80
+
81
+ model = RFDETRXLarge(accept_platform_model_license=True)
82
+ model.train(dataset_dir="path/to/dataset", epochs=50, lr=1e-4)
83
+ ```
84
+
85
+ ## Documentation
86
+
87
+ Visit the [RF-DETR documentation website](https://rfdetr.roboflow.com) to learn more about training, export, deployment, and the full model lineup.
88
+
89
+ ## License
90
+
91
+ RF-DETR+ code and model checkpoints are licensed under the Platform Model License 1.0 (PML-1.0). See [`LICENSE`](LICENSE) for details. These models require a [Roboflow](https://roboflow.com) account to run and fine-tune.
92
+
93
+ The core RF-DETR models (Nano through Large) are available under the Apache License 2.0 in the [`rfdetr`](https://github.com/roboflow/rf-detr) package.
94
+
95
+ ## Acknowledgements
96
+
97
+ Our work is built upon [LW-DETR](https://arxiv.org/pdf/2406.03459), [DINOv2](https://arxiv.org/pdf/2304.07193), and [Deformable DETR](https://arxiv.org/pdf/2010.04159). Thanks to their authors for their excellent work!
98
+
99
+ ## Citation
100
+
101
+ If you find our work helpful for your research, please consider citing the following BibTeX entry.
102
+
103
+ ```bibtex
104
+ @misc{rf-detr,
105
+ title={RF-DETR: Neural Architecture Search for Real-Time Detection Transformers},
106
+ author={Isaac Robinson and Peter Robicheaux and Matvei Popov and Deva Ramanan and Neehar Peri},
107
+ year={2025},
108
+ eprint={2511.09554},
109
+ archivePrefix={arXiv},
110
+ primaryClass={cs.CV},
111
+ url={https://arxiv.org/abs/2511.09554},
112
+ }
113
+ ```
114
+
115
+ ## Contribute
116
+
117
+ We welcome and appreciate all contributions! If you notice any issues or bugs, have questions, or would like to suggest new features, please [open an issue](https://github.com/roboflow/rf-detr-plus/issues/new) or pull request. By sharing your ideas and improvements, you help make RF-DETR better for everyone.
118
+
119
+ <p align="center">
120
+ <a href="https://youtube.com/roboflow"><img src="https://media.roboflow.com/notebooks/template/icons/purple/youtube.png?ik-sdk-version=javascript-1.4.3&updatedAt=1672949634652" width="3%"/></a>
121
+ <img src="https://raw.githubusercontent.com/ultralytics/assets/main/social/logo-transparent.png" width="3%"/>
122
+ <a href="https://roboflow.com"><img src="https://media.roboflow.com/notebooks/template/icons/purple/roboflow-app.png?ik-sdk-version=javascript-1.4.3&updatedAt=1672949746649" width="3%"/></a>
123
+ <img src="https://raw.githubusercontent.com/ultralytics/assets/main/social/logo-transparent.png" width="3%"/>
124
+ <a href="https://www.linkedin.com/company/roboflow-ai/"><img src="https://media.roboflow.com/notebooks/template/icons/purple/linkedin.png?ik-sdk-version=javascript-1.4.3&updatedAt=1672949633691" width="3%"/></a>
125
+ <img src="https://raw.githubusercontent.com/ultralytics/assets/main/social/logo-transparent.png" width="3%"/>
126
+ <a href="https://docs.roboflow.com"><img src="https://media.roboflow.com/notebooks/template/icons/purple/knowledge.png?ik-sdk-version=javascript-1.4.3&updatedAt=1672949634511" width="3%"/></a>
127
+ <img src="https://raw.githubusercontent.com/ultralytics/assets/main/social/logo-transparent.png" width="3%"/>
128
+ <a href="https://discuss.roboflow.com"><img src="https://media.roboflow.com/notebooks/template/icons/purple/forum.png?ik-sdk-version=javascript-1.4.3&updatedAt=1672949633584" width="3%"/></a>
129
+ <img src="https://raw.githubusercontent.com/ultralytics/assets/main/social/logo-transparent.png" width="3%"/>
130
+ <a href="https://blog.roboflow.com"><img src="https://media.roboflow.com/notebooks/template/icons/purple/blog.png?ik-sdk-version=javascript-1.4.3&updatedAt=1672949633605" width="3%"/></a>
131
+ </p>
@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
4
4
 
5
5
  [project]
6
6
  name = "rfdetr-plus"
7
- version = "0.0.0.dev0"
7
+ dynamic = ["version"]
8
8
  description = "RF-DETR+ Extension Package"
9
9
  readme = "README.md"
10
10
  authors = [
@@ -35,7 +35,7 @@ classifiers = [
35
35
  keywords = ["machine-learning", "deep-learning", "vision", "ML", "DL", "AI", "DETR", "RF-DETR", "Roboflow"]
36
36
 
37
37
  dependencies = [
38
- "rfdetr",
38
+ "rfdetr>=1.4.1,<2",
39
39
  ]
40
40
 
41
41
  [project.optional-dependencies]
@@ -52,27 +52,19 @@ tests = [
52
52
  "pytest-cov>=4,<8",
53
53
  ]
54
54
 
55
- docs = [
56
- "mkdocs-material[imaging]>=9.7",
57
- "mkdocstrings>=0.25.2,<0.30.0",
58
- "mkdocstrings-python>=1.10.9",
59
- "mike>=2.0.0",
60
- "mkdocs-jupyter>=0.24.3",
61
- "mkdocs-git-committers-plugin-2>=2.4.1; python_version >= '3.8' and python_version < '4'",
62
- "mkdocs-git-revision-date-localized-plugin>=1.2.4"
63
- ]
64
-
65
55
  dev = [
66
- "pre-commit",
56
+ "pre-commit>=4.5.1,<5",
67
57
  ]
68
58
 
69
59
  [project.urls]
70
60
  Homepage = "https://github.com/roboflow/rf-detr-plus"
71
61
 
72
62
  [tool.setuptools]
73
- package-dir = {"" = "src"}
74
63
  include-package-data = false
75
64
 
65
+ [tool.setuptools.dynamic]
66
+ version = {attr = "rfdetr_plus.__version__"}
67
+
76
68
  [tool.setuptools.packages.find]
77
69
  where = ["src"]
78
70
  include = ["rfdetr_plus*"]
@@ -0,0 +1,24 @@
1
+ # ------------------------------------------------------------------------
2
+ # RF-DETR+
3
+ # Copyright (c) 2026 Roboflow, Inc. All Rights Reserved.
4
+ # Licensed under the Platform Model License 1.0 [see LICENSE for details]
5
+ # ------------------------------------------------------------------------
6
+ """RF-DETR+ extension package.
7
+
8
+ Platform-licensed large-scale detection models (XLarge and 2XLarge)
9
+ for the RF-DETR family.
10
+ """
11
+
12
+ __version__ = "0.0.1"
13
+ __author__ = "Roboflow, Inc"
14
+ __author_email__ = "peter@roboflow.com"
15
+ __license__ = "PML-1.0"
16
+ __url__ = "https://github.com/roboflow/rf-detr-plus"
17
+ __docs__ = "https://rfdetr.roboflow.com"
18
+
19
+ from rfdetr_plus.models import RFDETR2XLarge, RFDETRXLarge
20
+
21
+ __all__ = [
22
+ "RFDETRXLarge",
23
+ "RFDETR2XLarge",
24
+ ]
@@ -3,3 +3,13 @@
3
3
  # Copyright (c) 2026 Roboflow, Inc. All Rights Reserved.
4
4
  # Licensed under the Platform Model License 1.0 [see LICENSE for details]
5
5
  # ------------------------------------------------------------------------
6
+ """RF-DETR+ model definitions."""
7
+
8
+ from rfdetr_plus.models.detection import RFDETR2XLarge, RFDETR2XLargeConfig, RFDETRXLarge, RFDETRXLargeConfig
9
+
10
+ __all__ = [
11
+ "RFDETRXLarge",
12
+ "RFDETRXLargeConfig",
13
+ "RFDETR2XLarge",
14
+ "RFDETR2XLargeConfig",
15
+ ]
@@ -1,12 +1,7 @@
1
1
  # ------------------------------------------------------------------------
2
- # Platform Model License 1.0 (PML-1.0)
2
+ # RF-DETR+
3
3
  # Copyright (c) 2026 Roboflow, Inc. All Rights Reserved.
4
- #
5
- # Licensed under the Platform Model License 1.0.
6
- # Use, modification, and distribution of code and checkpoints require
7
- # an active Roboflow platform plan or agreement.
8
- #
9
- # See the LICENSE.platform file for full terms and conditions.
4
+ # Licensed under the Platform Model License 1.0 [see LICENSE for details]
10
5
  # ------------------------------------------------------------------------
11
6
 
12
7
  from typing import Any, Literal
@@ -1,15 +1,10 @@
1
1
  # ------------------------------------------------------------------------
2
- # Platform Model License 1.0 (PML-1.0)
2
+ # RF-DETR+
3
3
  # Copyright (c) 2026 Roboflow, Inc. All Rights Reserved.
4
- #
5
- # Licensed under the Platform Model License 1.0.
6
- # Use, modification, and distribution of code and checkpoints require
7
- # an active Roboflow platform plan or agreement.
8
- #
9
- # See the LICENSE.platform file for full terms and conditions.
4
+ # Licensed under the Platform Model License 1.0 [see LICENSE for details]
10
5
  # ------------------------------------------------------------------------
11
6
 
12
- _PLATFORM_MODELS = {
7
+ PLATFORM_MODELS = {
13
8
  "rf-detr-xlarge.pth": "https://storage.googleapis.com/rfdetr/platform-licensed/rf-detr-xlarge.pth",
14
9
  "rf-detr-xxlarge.pth": "https://storage.googleapis.com/rfdetr/platform-licensed/rf-detr-xxlarge.pth",
15
10
  }
@@ -0,0 +1,162 @@
1
+ Metadata-Version: 2.4
2
+ Name: rfdetr-plus
3
+ Version: 0.0.1
4
+ Summary: RF-DETR+ Extension Package
5
+ Author-email: "Roboflow, Inc" <peter@roboflow.com>
6
+ License-Expression: LicenseRef-PML-1.0
7
+ Project-URL: Homepage, https://github.com/roboflow/rf-detr-plus
8
+ Keywords: machine-learning,deep-learning,vision,ML,DL,AI,DETR,RF-DETR,Roboflow
9
+ Classifier: Development Status :: 3 - Alpha
10
+ Classifier: Intended Audience :: Developers
11
+ Classifier: Intended Audience :: Education
12
+ Classifier: Intended Audience :: Science/Research
13
+ Classifier: Programming Language :: Python :: 3
14
+ Classifier: Programming Language :: Python :: 3.10
15
+ Classifier: Programming Language :: Python :: 3.11
16
+ Classifier: Programming Language :: Python :: 3.12
17
+ Classifier: Programming Language :: Python :: 3.13
18
+ Classifier: Programming Language :: Python :: 3 :: Only
19
+ Classifier: Topic :: Software Development
20
+ Classifier: Topic :: Scientific/Engineering
21
+ Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
22
+ Classifier: Typing :: Typed
23
+ Classifier: Operating System :: POSIX
24
+ Classifier: Operating System :: Unix
25
+ Classifier: Operating System :: MacOS
26
+ Requires-Python: >=3.10
27
+ Description-Content-Type: text/markdown
28
+ License-File: LICENSE
29
+ Requires-Dist: rfdetr<2,>=1.4.1
30
+ Dynamic: license-file
31
+
32
+ # RF-DETR+: Large-Scale Detection Models for RF-DETR
33
+
34
+ [![version](https://badge.fury.io/py/rfdetr-plus.svg)](https://badge.fury.io/py/rfdetr-plus)
35
+ [![downloads](https://img.shields.io/pypi/dm/rfdetr-plus)](https://pypistats.org/packages/rfdetr-plus)
36
+ [![python-version](https://img.shields.io/pypi/pyversions/rfdetr-plus)](https://badge.fury.io/py/rfdetr-plus)
37
+ [![license](https://img.shields.io/badge/license-PML--1.0-blue)](https://github.com/roboflow/rf-detr-plus/blob/main/LICENSE)
38
+
39
+ [![discord](https://img.shields.io/discord/1159501506232451173?logo=discord&label=discord&labelColor=fff&color=5865f2&link=https%3A%2F%2Fdiscord.gg%2FGbfgXGJ8Bk)](https://discord.gg/GbfgXGJ8Bk)
40
+
41
+ RF-DETR+ is the extension package for [RF-DETR](https://github.com/roboflow/rf-detr) that provides the **XLarge** and **2XLarge** detection models. These are the largest and most accurate models in the RF-DETR family, pushing state-of-the-art accuracy on [Microsoft COCO](https://cocodataset.org/#home) and [RF100-VL](https://github.com/roboflow/rf100-vl) while retaining real-time inference speeds.
42
+
43
+ RF-DETR+ models use a DINOv2 vision transformer backbone at higher resolutions and larger feature dimensions than the core RF-DETR lineup, unlocking top-tier detection accuracy for applications where precision matters most.
44
+
45
+ ## Install
46
+
47
+ Install RF-DETR+ in a [**Python>=3.10**](https://www.python.org/) environment with `pip`. This will also install [`rfdetr`](https://github.com/roboflow/rf-detr) as a dependency.
48
+
49
+ ```bash
50
+ pip install rfdetr-plus
51
+ ```
52
+
53
+ <details>
54
+ <summary>Install from source</summary>
55
+
56
+ <br>
57
+
58
+ ```bash
59
+ pip install git+https://github.com/roboflow/rf-detr-plus.git
60
+ ```
61
+
62
+ </details>
63
+
64
+ ## Benchmarks
65
+
66
+ RF-DETR+ XLarge and 2XLarge sit at the top of the RF-DETR accuracy/latency curve, delivering the highest COCO AP scores in the family. All latency numbers were measured on an NVIDIA T4 using TensorRT, FP16, and batch size 1.
67
+
68
+ | Size | Class | COCO AP<sub>50</sub> | COCO AP<sub>50:95</sub> | RF100VL AP<sub>50</sub> | RF100VL AP<sub>50:95</sub> | Latency (ms) | Params (M) | Resolution | Package / License |
69
+ | :----: | :-------------: | :------------------: | :---------------------: | :---------------------: | :------------------------: | :----------: | :--------: | :--------: | :----------------------------------------------------------: |
70
+ | N | `RFDETRNano` | 67.6 | 48.4 | 85.0 | 57.7 | 2.3 | 30.5 | 384x384 | [`rfdetr`](https://github.com/roboflow/rf-detr) / Apache 2.0 |
71
+ | S | `RFDETRSmall` | 72.1 | 53.0 | 86.7 | 60.2 | 3.5 | 32.1 | 512x512 | [`rfdetr`](https://github.com/roboflow/rf-detr) / Apache 2.0 |
72
+ | M | `RFDETRMedium` | 73.6 | 54.7 | 87.4 | 61.2 | 4.4 | 33.7 | 576x576 | [`rfdetr`](https://github.com/roboflow/rf-detr) / Apache 2.0 |
73
+ | L | `RFDETRLarge` | 75.1 | 56.5 | 88.2 | 62.2 | 6.8 | 33.9 | 704x704 | [`rfdetr`](https://github.com/roboflow/rf-detr) / Apache 2.0 |
74
+ | ⭐ XL | `RFDETRXLarge` | 77.4 | 58.6 | 88.5 | 62.9 | 11.5 | 126.4 | 700x700 | `rfdetr_plus` / [PML 1.0](LICENSE) |
75
+ | ⭐ 2XL | `RFDETR2XLarge` | 78.5 | 60.1 | 89.0 | 63.2 | 17.2 | 126.9 | 880x880 | `rfdetr_plus` / [PML 1.0](LICENSE) |
76
+
77
+ ## Run Models
78
+
79
+ Install with the `plus` extra to get XL and 2XL models alongside core RF-DETR:
80
+
81
+ ```bash
82
+ pip install rfdetr_plus
83
+ ```
84
+
85
+ RF-DETR+ models require you to accept the Platform Model License before use. Once accepted, usage mirrors the standard RF-DETR API -- you import directly from `rfdetr`:
86
+
87
+ ```python
88
+ import requests
89
+ import supervision as sv
90
+ from PIL import Image
91
+ from rfdetr import RFDETRXLarge
92
+ from rfdetr.util.coco_classes import COCO_CLASSES
93
+
94
+ model = RFDETRXLarge(accept_platform_model_license=True)
95
+
96
+ image = Image.open(requests.get("https://media.roboflow.com/dog.jpg", stream=True).raw)
97
+ detections = model.predict(image, threshold=0.5)
98
+
99
+ labels = [f"{COCO_CLASSES[class_id]}" for class_id in detections.class_id]
100
+
101
+ annotated_image = sv.BoxAnnotator().annotate(image, detections)
102
+ annotated_image = sv.LabelAnnotator().annotate(annotated_image, detections, labels)
103
+ ```
104
+
105
+ ### Train Models
106
+
107
+ RF-DETR+ models support fine-tuning with the same training API as core RF-DETR. You can train on your own dataset or use datasets from [Roboflow Universe](https://universe.roboflow.com/).
108
+
109
+ ```python
110
+ from rfdetr import RFDETRXLarge
111
+
112
+ model = RFDETRXLarge(accept_platform_model_license=True)
113
+ model.train(dataset_dir="path/to/dataset", epochs=50, lr=1e-4)
114
+ ```
115
+
116
+ ## Documentation
117
+
118
+ Visit the [RF-DETR documentation website](https://rfdetr.roboflow.com) to learn more about training, export, deployment, and the full model lineup.
119
+
120
+ ## License
121
+
122
+ RF-DETR+ code and model checkpoints are licensed under the Platform Model License 1.0 (PML-1.0). See [`LICENSE`](LICENSE) for details. These models require a [Roboflow](https://roboflow.com) account to run and fine-tune.
123
+
124
+ The core RF-DETR models (Nano through Large) are available under the Apache License 2.0 in the [`rfdetr`](https://github.com/roboflow/rf-detr) package.
125
+
126
+ ## Acknowledgements
127
+
128
+ Our work is built upon [LW-DETR](https://arxiv.org/pdf/2406.03459), [DINOv2](https://arxiv.org/pdf/2304.07193), and [Deformable DETR](https://arxiv.org/pdf/2010.04159). Thanks to their authors for their excellent work!
129
+
130
+ ## Citation
131
+
132
+ If you find our work helpful for your research, please consider citing the following BibTeX entry.
133
+
134
+ ```bibtex
135
+ @misc{rf-detr,
136
+ title={RF-DETR: Neural Architecture Search for Real-Time Detection Transformers},
137
+ author={Isaac Robinson and Peter Robicheaux and Matvei Popov and Deva Ramanan and Neehar Peri},
138
+ year={2025},
139
+ eprint={2511.09554},
140
+ archivePrefix={arXiv},
141
+ primaryClass={cs.CV},
142
+ url={https://arxiv.org/abs/2511.09554},
143
+ }
144
+ ```
145
+
146
+ ## Contribute
147
+
148
+ We welcome and appreciate all contributions! If you notice any issues or bugs, have questions, or would like to suggest new features, please [open an issue](https://github.com/roboflow/rf-detr-plus/issues/new) or pull request. By sharing your ideas and improvements, you help make RF-DETR better for everyone.
149
+
150
+ <p align="center">
151
+ <a href="https://youtube.com/roboflow"><img src="https://media.roboflow.com/notebooks/template/icons/purple/youtube.png?ik-sdk-version=javascript-1.4.3&updatedAt=1672949634652" width="3%"/></a>
152
+ <img src="https://raw.githubusercontent.com/ultralytics/assets/main/social/logo-transparent.png" width="3%"/>
153
+ <a href="https://roboflow.com"><img src="https://media.roboflow.com/notebooks/template/icons/purple/roboflow-app.png?ik-sdk-version=javascript-1.4.3&updatedAt=1672949746649" width="3%"/></a>
154
+ <img src="https://raw.githubusercontent.com/ultralytics/assets/main/social/logo-transparent.png" width="3%"/>
155
+ <a href="https://www.linkedin.com/company/roboflow-ai/"><img src="https://media.roboflow.com/notebooks/template/icons/purple/linkedin.png?ik-sdk-version=javascript-1.4.3&updatedAt=1672949633691" width="3%"/></a>
156
+ <img src="https://raw.githubusercontent.com/ultralytics/assets/main/social/logo-transparent.png" width="3%"/>
157
+ <a href="https://docs.roboflow.com"><img src="https://media.roboflow.com/notebooks/template/icons/purple/knowledge.png?ik-sdk-version=javascript-1.4.3&updatedAt=1672949634511" width="3%"/></a>
158
+ <img src="https://raw.githubusercontent.com/ultralytics/assets/main/social/logo-transparent.png" width="3%"/>
159
+ <a href="https://discuss.roboflow.com"><img src="https://media.roboflow.com/notebooks/template/icons/purple/forum.png?ik-sdk-version=javascript-1.4.3&updatedAt=1672949633584" width="3%"/></a>
160
+ <img src="https://raw.githubusercontent.com/ultralytics/assets/main/social/logo-transparent.png" width="3%"/>
161
+ <a href="https://blog.roboflow.com"><img src="https://media.roboflow.com/notebooks/template/icons/purple/blog.png?ik-sdk-version=javascript-1.4.3&updatedAt=1672949633605" width="3%"/></a>
162
+ </p>
@@ -0,0 +1 @@
1
+ rfdetr<2,>=1.4.1
@@ -1,98 +0,0 @@
1
- Metadata-Version: 2.4
2
- Name: rfdetr-plus
3
- Version: 0.0.0.dev0
4
- Summary: RF-DETR+ Extension Package
5
- Author-email: "Roboflow, Inc" <peter@roboflow.com>
6
- License-Expression: LicenseRef-PML-1.0
7
- Project-URL: Homepage, https://github.com/roboflow/rf-detr-plus
8
- Keywords: machine-learning,deep-learning,vision,ML,DL,AI,DETR,RF-DETR,Roboflow
9
- Classifier: Development Status :: 3 - Alpha
10
- Classifier: Intended Audience :: Developers
11
- Classifier: Intended Audience :: Education
12
- Classifier: Intended Audience :: Science/Research
13
- Classifier: Programming Language :: Python :: 3
14
- Classifier: Programming Language :: Python :: 3.10
15
- Classifier: Programming Language :: Python :: 3.11
16
- Classifier: Programming Language :: Python :: 3.12
17
- Classifier: Programming Language :: Python :: 3.13
18
- Classifier: Programming Language :: Python :: 3 :: Only
19
- Classifier: Topic :: Software Development
20
- Classifier: Topic :: Scientific/Engineering
21
- Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
22
- Classifier: Typing :: Typed
23
- Classifier: Operating System :: POSIX
24
- Classifier: Operating System :: Unix
25
- Classifier: Operating System :: MacOS
26
- Requires-Python: >=3.10
27
- Description-Content-Type: text/markdown
28
- License-File: LICENSE
29
- Requires-Dist: rfdetr
30
- Dynamic: license-file
31
-
32
- # RF-DETR+
33
-
34
- Extension package for [RF-DETR](https://github.com/roboflow/rf-detr) containing Platform Model License (PML) code.
35
-
36
- ## Installation
37
-
38
- ```bash
39
- pip install rfdetr-plus
40
- ```
41
-
42
- Or for development:
43
-
44
- ```bash
45
- pip install -e ".[tests]"
46
- ```
47
-
48
- ## Structure
49
-
50
- ```
51
- rf-detr-plus/
52
- ├── .github/
53
- │ ├── workflows/ # CI/CD pipelines
54
- │ └── ISSUE_TEMPLATE/ # Issue templates
55
- ├── src/
56
- │ └── rfdetr_plus/ # Main package
57
- │ ├── __init__.py
58
- │ └── py.typed
59
- ├── tests/ # Test suite
60
- ├── .codecov.yml
61
- ├── .gitignore
62
- ├── .pre-commit-config.yaml
63
- ├── CONTRIBUTING.md
64
- ├── LICENSE
65
- ├── README.md
66
- └── pyproject.toml
67
- ```
68
-
69
- ## Development
70
-
71
- ### Code Quality
72
-
73
- We use `pre-commit` to ensure code quality:
74
-
75
- ```bash
76
- pre-commit install
77
- pre-commit run --all-files
78
- ```
79
-
80
- ### Tests
81
-
82
- [`pytest`](https://docs.pytest.org/) is used to run tests:
83
-
84
- ```bash
85
- pytest tests/ -v
86
- ```
87
-
88
- ### License Headers
89
-
90
- All Python files must include the following license header:
91
-
92
- ```python
93
- # ------------------------------------------------------------------------
94
- # RF-DETR+
95
- # Copyright (c) 2026 Roboflow, Inc. All Rights Reserved.
96
- # Licensed under the Platform Model License 1.0 [see LICENSE for details]
97
- # ------------------------------------------------------------------------
98
- ```
@@ -1,67 +0,0 @@
1
- # RF-DETR+
2
-
3
- Extension package for [RF-DETR](https://github.com/roboflow/rf-detr) containing Platform Model License (PML) code.
4
-
5
- ## Installation
6
-
7
- ```bash
8
- pip install rfdetr-plus
9
- ```
10
-
11
- Or for development:
12
-
13
- ```bash
14
- pip install -e ".[tests]"
15
- ```
16
-
17
- ## Structure
18
-
19
- ```
20
- rf-detr-plus/
21
- ├── .github/
22
- │ ├── workflows/ # CI/CD pipelines
23
- │ └── ISSUE_TEMPLATE/ # Issue templates
24
- ├── src/
25
- │ └── rfdetr_plus/ # Main package
26
- │ ├── __init__.py
27
- │ └── py.typed
28
- ├── tests/ # Test suite
29
- ├── .codecov.yml
30
- ├── .gitignore
31
- ├── .pre-commit-config.yaml
32
- ├── CONTRIBUTING.md
33
- ├── LICENSE
34
- ├── README.md
35
- └── pyproject.toml
36
- ```
37
-
38
- ## Development
39
-
40
- ### Code Quality
41
-
42
- We use `pre-commit` to ensure code quality:
43
-
44
- ```bash
45
- pre-commit install
46
- pre-commit run --all-files
47
- ```
48
-
49
- ### Tests
50
-
51
- [`pytest`](https://docs.pytest.org/) is used to run tests:
52
-
53
- ```bash
54
- pytest tests/ -v
55
- ```
56
-
57
- ### License Headers
58
-
59
- All Python files must include the following license header:
60
-
61
- ```python
62
- # ------------------------------------------------------------------------
63
- # RF-DETR+
64
- # Copyright (c) 2026 Roboflow, Inc. All Rights Reserved.
65
- # Licensed under the Platform Model License 1.0 [see LICENSE for details]
66
- # ------------------------------------------------------------------------
67
- ```
File without changes
@@ -1,98 +0,0 @@
1
- Metadata-Version: 2.4
2
- Name: rfdetr-plus
3
- Version: 0.0.0.dev0
4
- Summary: RF-DETR+ Extension Package
5
- Author-email: "Roboflow, Inc" <peter@roboflow.com>
6
- License-Expression: LicenseRef-PML-1.0
7
- Project-URL: Homepage, https://github.com/roboflow/rf-detr-plus
8
- Keywords: machine-learning,deep-learning,vision,ML,DL,AI,DETR,RF-DETR,Roboflow
9
- Classifier: Development Status :: 3 - Alpha
10
- Classifier: Intended Audience :: Developers
11
- Classifier: Intended Audience :: Education
12
- Classifier: Intended Audience :: Science/Research
13
- Classifier: Programming Language :: Python :: 3
14
- Classifier: Programming Language :: Python :: 3.10
15
- Classifier: Programming Language :: Python :: 3.11
16
- Classifier: Programming Language :: Python :: 3.12
17
- Classifier: Programming Language :: Python :: 3.13
18
- Classifier: Programming Language :: Python :: 3 :: Only
19
- Classifier: Topic :: Software Development
20
- Classifier: Topic :: Scientific/Engineering
21
- Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
22
- Classifier: Typing :: Typed
23
- Classifier: Operating System :: POSIX
24
- Classifier: Operating System :: Unix
25
- Classifier: Operating System :: MacOS
26
- Requires-Python: >=3.10
27
- Description-Content-Type: text/markdown
28
- License-File: LICENSE
29
- Requires-Dist: rfdetr
30
- Dynamic: license-file
31
-
32
- # RF-DETR+
33
-
34
- Extension package for [RF-DETR](https://github.com/roboflow/rf-detr) containing Platform Model License (PML) code.
35
-
36
- ## Installation
37
-
38
- ```bash
39
- pip install rfdetr-plus
40
- ```
41
-
42
- Or for development:
43
-
44
- ```bash
45
- pip install -e ".[tests]"
46
- ```
47
-
48
- ## Structure
49
-
50
- ```
51
- rf-detr-plus/
52
- ├── .github/
53
- │ ├── workflows/ # CI/CD pipelines
54
- │ └── ISSUE_TEMPLATE/ # Issue templates
55
- ├── src/
56
- │ └── rfdetr_plus/ # Main package
57
- │ ├── __init__.py
58
- │ └── py.typed
59
- ├── tests/ # Test suite
60
- ├── .codecov.yml
61
- ├── .gitignore
62
- ├── .pre-commit-config.yaml
63
- ├── CONTRIBUTING.md
64
- ├── LICENSE
65
- ├── README.md
66
- └── pyproject.toml
67
- ```
68
-
69
- ## Development
70
-
71
- ### Code Quality
72
-
73
- We use `pre-commit` to ensure code quality:
74
-
75
- ```bash
76
- pre-commit install
77
- pre-commit run --all-files
78
- ```
79
-
80
- ### Tests
81
-
82
- [`pytest`](https://docs.pytest.org/) is used to run tests:
83
-
84
- ```bash
85
- pytest tests/ -v
86
- ```
87
-
88
- ### License Headers
89
-
90
- All Python files must include the following license header:
91
-
92
- ```python
93
- # ------------------------------------------------------------------------
94
- # RF-DETR+
95
- # Copyright (c) 2026 Roboflow, Inc. All Rights Reserved.
96
- # Licensed under the Platform Model License 1.0 [see LICENSE for details]
97
- # ------------------------------------------------------------------------
98
- ```
File without changes
File without changes