pydfine 0.0.1__tar.gz

pydfine-0.0.1/LICENSE

@@ -0,0 +1,201 @@
+                                 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 [yyyy] [name of copyright owner]
+
+   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.

pydfine-0.0.1/NOTICE

@@ -0,0 +1,21 @@
+pydfine
+Copyright 2026 pydfine contributors
+
+This product is licensed under the Apache License, Version 2.0 (see LICENSE).
+
+It is a derivative work of D-FINE and includes source code ported from it:
+
+  D-FINE — https://github.com/Peterande/D-FINE
+  Copyright 2024 The D-FINE Authors
+  Licensed under the Apache License, Version 2.0.
+
+The ported modules under `dfine/backends/native/` preserve upstream layer and
+parameter names (so released D-FINE `.pth` checkpoints load unchanged) and each
+carries a per-file header naming its source file, license, and the changes made.
+
+D-FINE itself builds on earlier work; those attributions are retained in the
+relevant ported source files:
+
+  - RT-DETR — Copyright 2023 lyuwenyu — Apache-2.0
+  - DETR    — Copyright 2020 Facebook, Inc. — Apache-2.0
+  - PaddleDetection / PP-HGNetV2 — Apache-2.0

pydfine-0.0.1/PKG-INFO

@@ -0,0 +1,239 @@
+Metadata-Version: 2.4
+Name: pydfine
+Version: 0.0.1
+Summary: Ultralytics-style, config-first Python library for the D-FINE real-time object detector.
+Author: pydfine contributors
+License: Apache-2.0
+Project-URL: Homepage, https://github.com/HoshiBatista/pydfine
+Project-URL: Documentation, https://hoshibatista.github.io/pydfine/
+Project-URL: Repository, https://github.com/HoshiBatista/pydfine
+Project-URL: Issues, https://github.com/HoshiBatista/pydfine/issues
+Project-URL: Upstream, https://github.com/Peterande/D-FINE
+Project-URL: Paper, https://arxiv.org/abs/2410.13842
+Keywords: d-fine,object-detection,detr,computer-vision,pytorch,onnx
+Classifier: Development Status :: 4 - Beta
+Classifier: Intended Audience :: Developers
+Classifier: Intended Audience :: Science/Research
+Classifier: License :: OSI Approved :: Apache Software License
+Classifier: Operating System :: OS Independent
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.9
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: 3.13
+Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
+Classifier: Topic :: Scientific/Engineering :: Image Recognition
+Classifier: Typing :: Typed
+Requires-Python: >=3.9
+Description-Content-Type: text/markdown
+License-File: LICENSE
+License-File: NOTICE
+Requires-Dist: numpy>=1.21
+Provides-Extra: torch
+Requires-Dist: torch>=2.0.1; extra == "torch"
+Requires-Dist: torchvision>=0.15.2; extra == "torch"
+Requires-Dist: pillow>=9.0; extra == "torch"
+Provides-Extra: train
+Requires-Dist: pydfine[torch]; extra == "train"
+Requires-Dist: scipy>=1.10; extra == "train"
+Requires-Dist: faster-coco-eval>=1.6.6; extra == "train"
+Requires-Dist: tensorboard>=2.13; extra == "train"
+Requires-Dist: matplotlib>=3.6; extra == "train"
+Requires-Dist: pyyaml>=6.0; extra == "train"
+Provides-Extra: export
+Requires-Dist: onnx>=1.14; extra == "export"
+Requires-Dist: onnxruntime>=1.16; extra == "export"
+Requires-Dist: onnxsim>=0.4; extra == "export"
+Provides-Extra: video
+Requires-Dist: pydfine[torch]; extra == "video"
+Requires-Dist: opencv-python>=4.6; extra == "video"
+Provides-Extra: interop
+Requires-Dist: pandas>=1.3; extra == "interop"
+Requires-Dist: supervision>=0.18; extra == "interop"
+Provides-Extra: track
+Requires-Dist: pydfine[video]; extra == "track"
+Requires-Dist: scipy>=1.10; extra == "track"
+Provides-Extra: docs
+Requires-Dist: mkdocs>=1.6; extra == "docs"
+Requires-Dist: mkdocs-material>=9.5; extra == "docs"
+Requires-Dist: mkdocstrings[python]>=0.26; extra == "docs"
+Provides-Extra: dev
+Requires-Dist: pydfine[torch]; extra == "dev"
+Requires-Dist: ruff==0.15.21; extra == "dev"
+Requires-Dist: pytest>=7.0; extra == "dev"
+Requires-Dist: pytest-cov>=4.0; extra == "dev"
+Requires-Dist: pre-commit>=3.5; extra == "dev"
+Requires-Dist: opencv-python-headless>=4.6; extra == "dev"
+Requires-Dist: scipy>=1.10; extra == "dev"
+Requires-Dist: faster-coco-eval>=1.6.6; extra == "dev"
+Requires-Dist: pyyaml>=6.0; extra == "dev"
+Requires-Dist: onnx>=1.14; extra == "dev"
+Requires-Dist: onnxruntime>=1.16; extra == "dev"
+Requires-Dist: pandas>=1.3; extra == "dev"
+Requires-Dist: supervision>=0.18; extra == "dev"
+Dynamic: license-file
+
+# dfine
+
+A batteries-included Python library for the **D-FINE** real-time object detector
+([Peterande/D-FINE](https://github.com/Peterande/D-FINE), ICLR 2025 Spotlight),
+with an `ultralytics`-style developer experience.
+
+[![PyPI](https://img.shields.io/pypi/v/pydfine.svg)](https://pypi.org/project/pydfine/)
+[![Python](https://img.shields.io/pypi/pyversions/pydfine.svg)](https://pypi.org/project/pydfine/)
+[![License](https://img.shields.io/badge/license-Apache--2.0-blue.svg)](LICENSE)
+
+📖 **Documentation:** <https://hoshibatista.github.io/pydfine/>
+
+Install: `pip install pydfine` (core, torch-free) — `import dfine` to use it.
+
+**Design goal:** the entire model — backbone, encoder, decoder, losses, denoising,
+training, augmentation — is configured through **typed Python parameters on one
+class**. No YAML files, no config-registry indirection, no `torchrun` incantations.
+
+```python
+from dfine import DFINE
+
+# Presets fill sensible defaults; every single field is overridable inline.
+model = DFINE(
+    size="l",                 # n | s | m | l | x  -> sets backbone, dims, depths
+    num_classes=80,
+    num_queries=300,
+    hidden_dim=256,
+    reg_max=32,               # Fine-grained Distribution Refinement bins
+    backbone="hgnetv2_b4",
+    backbone_pretrained=True,
+    device="cuda",
+)
+
+results = model.predict("street.jpg", conf=0.4)
+results[0].save("out.jpg")
+
+model.train(data="dataset/", epochs=72, imgsz=640, batch=32)
+metrics = model.val()
+model.export(format="onnx")
+```
+
+Fully custom architecture, no preset:
+
+```python
+model = DFINE(
+    num_classes=3,
+    backbone="hgnetv2_b0", use_lab=True, freeze_at=-1,
+    hidden_dim=256, encoder_dim_feedforward=1024, encoder_layers=1, nhead=8,
+    decoder_layers=4, eval_idx=-1, num_levels=3, num_points=[3, 6, 3],
+    reg_max=32, reg_scale=4.0, lqe_layers=2,
+    num_denoising=100, label_noise_ratio=0.5, box_noise_scale=1.0,
+    class_names=["cat", "dog", "bird"],
+)
+```
+
+## Status
+
+Inference is complete and bit-exact with upstream; the training stack (loop, data,
+augmentation, COCO `val`, multi-GPU DDP, visualization) is in — only `export` remains.
+Done so far:
+
+- **Config-first core** — `DFINEConfig` (every model/training param as a typed field),
+  verified `n/s/m/l/x` presets, validation, checkpoint registry, `dfine models` CLI.
+- **Native model port (Path A)** under `dfine/backends/native/` — the full
+  **backbone → encoder → decoder** stack ported from upstream `src/` with the
+  YAML/registry layer stripped: `HGNetv2`, `HybridEncoder`, and `DFINETransformer`
+  (FDR head, LQE, contrastive denoising). Layer/param names preserved so released
+  `.pth` load unchanged. Each module builds from the config via `from_config(cfg)`.
+
+- **Working inference** — assembled `DFINE` model + `DFINEPostProcessor`, upstream
+  `.pth` loading (`registry`/`downloads`, `from_pretrained`), and the public
+  `DFINE(...).predict(...) -> Results` API (`.boxes.xyxy/.conf/.cls`, `.plot()/.save()`).
+
+- **Video** — `DFINE.predict_video(source, output=...)` writes an annotated mp4, or
+  `stream=True` yields per-frame `Results` (needs `pip install pydfine[video]`).
+
+- **Training loss** — `HungarianMatcher` + `DFINECriterion` (VFL + L1 + GIoU + FGL +
+  DDF) ported and wired from the config; consumes the decoder's training-mode output.
+
+- **Training loop** — `DFINE.train(train_loader, epochs=...)` runs the ported D-FINE
+  loop (AdamW param groups, EMA, AMP, grad clip, warmup + flat-cosine LR) with the same
+  **progress visualization as upstream**: a live console readout (`MetricLogger`) plus
+  TensorBoard scalars and a `loss_curve.png` under `output_dir` (needs
+  `pip install pydfine[train]`; W&B optional).
+
+- **COCO data + augmentation** — `dfine.train.dataset.build_coco_dataloader(img_folder,
+  ann_file, cfg=...)` gives a ready `(images, targets)` loader (contiguous-label remap,
+  multi-scale collate); pass `transforms=dfine.train.augment.train_transforms(imgsz,
+  stop_epoch=...)` for D-FINE's full augment pipeline (photometric distort, zoom-out,
+  IoU-crop, H-flip) with the two-phase no-aug tail. Feeds straight into `DFINE.train`.
+
+- **Have a YOLO dataset?** Convert it once — `dfine convert path/to/yolo path/to/coco`
+  (or `dfine.yolo_to_coco(...)`) — then `DFINE.train(data="path/to/coco")`. It reads the
+  `images/<split>` + `labels/<split>` layout (and `data.yaml` names) and writes the COCO
+  layout with 0-indexed categories that line up with the model's labels.
+
+Only `export` (Phase 3) remains. See [`docs/ROADMAP.md`](docs/ROADMAP.md) for the full
+status.
+
+```python
+from dfine import DFINEConfig
+
+cfg = DFINEConfig.preset("l", num_classes=3)   # verified upstream defaults
+cfg = DFINEConfig.preset("n")                  # 2-level, hidden_dim=128
+```
+
+The ported modules already run end-to-end (needs the `torch` extra installed):
+
+```python
+import torch
+from dfine import DFINEConfig
+from dfine.backends.native import HGNetv2, HybridEncoder, DFINETransformer
+
+cfg = DFINEConfig.preset("l", num_classes=80)
+backbone = HGNetv2.from_config(cfg).eval()
+encoder = HybridEncoder.from_config(cfg).eval()
+decoder = DFINETransformer.from_config(cfg).eval()
+
+out = decoder(encoder(backbone(torch.randn(1, 3, cfg.imgsz, cfg.imgsz))))
+# out["pred_logits"]: (1, 300, 80)   out["pred_boxes"]: (1, 300, 4)  [cxcywh, 0..1]
+```
+
+> The one-class `DFINE(...)` façade at the top of this README works today for
+> **inference** (`predict`/`load`/`from_pretrained`), **training** — both
+> `train(data="coco/", epochs=...)` (a standard COCO root; the loaders are built for
+> you) and `train(train_loader, epochs=...)` (a hand-built loader) — and **COCO
+> evaluation** (`val(data="coco/")` → the 12 named COCO metrics, also run each epoch
+> during `train`), all with the `pydfine[train]` extra. **Multi-GPU** is a single kwarg:
+> `train(data="coco/", devices=N)` spawns one DDP worker per GPU (or launch with
+> `torchrun` and call `train(...)` as usual). `export` is phase-stubbed and raises a
+> clear "arriving in Phase N" until implemented.
+
+## Why this exists
+
+Upstream D-FINE is an excellent research repo, but using it means editing YAML,
+copying config include-trees, and launching scripts. This library turns all of that
+into one importable, fully-typed class with presets — so a developer can go from
+`pip install` to a trained custom detector without touching a config file.
+
+## For contributors and AI agents
+
+This project is built to be developed largely by coding agents (Claude Code / any
+agent that reads `AGENTS.md`). Start here:
+
+| File | Purpose |
+|---|---|
+| [`AGENTS.md`](AGENTS.md) | **Canonical agent guide** — architecture, conventions, workflow, commands, definition of done. Read first. |
+| [`CLAUDE.md`](CLAUDE.md) | Claude Code–specific notes; defers to `AGENTS.md`. |
+| [`docs/ARCHITECTURE.md`](docs/ARCHITECTURE.md) | How D-FINE works and how we re-shape it into Python. |
+| [`docs/CONFIG_REFERENCE.md`](docs/CONFIG_REFERENCE.md) | Every model parameter, default, and per-size preset. The heart of the "one class, many params" design. |
+| [`docs/ROADMAP.md`](docs/ROADMAP.md) | Phased, checkbox task plan. |
+
+## License & attribution
+
+pydfine is licensed under the [Apache License 2.0](LICENSE).
+
+It is a **derivative work of D-FINE**
+([Peterande/D-FINE](https://github.com/Peterande/D-FINE), Apache-2.0, © 2024 The
+D-FINE Authors): the model is a native port of upstream `src/`, with layer and
+parameter names preserved so released `.pth` checkpoints load unchanged. Every ported
+module under `dfine/backends/native/` carries a per-file header crediting its source
+and describing the changes. See [`NOTICE`](NOTICE) for the full attribution, including
+D-FINE's own lineage (RT-DETR, DETR, PaddleDetection).

pydfine-0.0.1/README.md

@@ -0,0 +1,163 @@
+# dfine
+
+A batteries-included Python library for the **D-FINE** real-time object detector
+([Peterande/D-FINE](https://github.com/Peterande/D-FINE), ICLR 2025 Spotlight),
+with an `ultralytics`-style developer experience.
+
+[![PyPI](https://img.shields.io/pypi/v/pydfine.svg)](https://pypi.org/project/pydfine/)
+[![Python](https://img.shields.io/pypi/pyversions/pydfine.svg)](https://pypi.org/project/pydfine/)
+[![License](https://img.shields.io/badge/license-Apache--2.0-blue.svg)](LICENSE)
+
+📖 **Documentation:** <https://hoshibatista.github.io/pydfine/>
+
+Install: `pip install pydfine` (core, torch-free) — `import dfine` to use it.
+
+**Design goal:** the entire model — backbone, encoder, decoder, losses, denoising,
+training, augmentation — is configured through **typed Python parameters on one
+class**. No YAML files, no config-registry indirection, no `torchrun` incantations.
+
+```python
+from dfine import DFINE
+
+# Presets fill sensible defaults; every single field is overridable inline.
+model = DFINE(
+    size="l",                 # n | s | m | l | x  -> sets backbone, dims, depths
+    num_classes=80,
+    num_queries=300,
+    hidden_dim=256,
+    reg_max=32,               # Fine-grained Distribution Refinement bins
+    backbone="hgnetv2_b4",
+    backbone_pretrained=True,
+    device="cuda",
+)
+
+results = model.predict("street.jpg", conf=0.4)
+results[0].save("out.jpg")
+
+model.train(data="dataset/", epochs=72, imgsz=640, batch=32)
+metrics = model.val()
+model.export(format="onnx")
+```
+
+Fully custom architecture, no preset:
+
+```python
+model = DFINE(
+    num_classes=3,
+    backbone="hgnetv2_b0", use_lab=True, freeze_at=-1,
+    hidden_dim=256, encoder_dim_feedforward=1024, encoder_layers=1, nhead=8,
+    decoder_layers=4, eval_idx=-1, num_levels=3, num_points=[3, 6, 3],
+    reg_max=32, reg_scale=4.0, lqe_layers=2,
+    num_denoising=100, label_noise_ratio=0.5, box_noise_scale=1.0,
+    class_names=["cat", "dog", "bird"],
+)
+```
+
+## Status
+
+Inference is complete and bit-exact with upstream; the training stack (loop, data,
+augmentation, COCO `val`, multi-GPU DDP, visualization) is in — only `export` remains.
+Done so far:
+
+- **Config-first core** — `DFINEConfig` (every model/training param as a typed field),
+  verified `n/s/m/l/x` presets, validation, checkpoint registry, `dfine models` CLI.
+- **Native model port (Path A)** under `dfine/backends/native/` — the full
+  **backbone → encoder → decoder** stack ported from upstream `src/` with the
+  YAML/registry layer stripped: `HGNetv2`, `HybridEncoder`, and `DFINETransformer`
+  (FDR head, LQE, contrastive denoising). Layer/param names preserved so released
+  `.pth` load unchanged. Each module builds from the config via `from_config(cfg)`.
+
+- **Working inference** — assembled `DFINE` model + `DFINEPostProcessor`, upstream
+  `.pth` loading (`registry`/`downloads`, `from_pretrained`), and the public
+  `DFINE(...).predict(...) -> Results` API (`.boxes.xyxy/.conf/.cls`, `.plot()/.save()`).
+
+- **Video** — `DFINE.predict_video(source, output=...)` writes an annotated mp4, or
+  `stream=True` yields per-frame `Results` (needs `pip install pydfine[video]`).
+
+- **Training loss** — `HungarianMatcher` + `DFINECriterion` (VFL + L1 + GIoU + FGL +
+  DDF) ported and wired from the config; consumes the decoder's training-mode output.
+
+- **Training loop** — `DFINE.train(train_loader, epochs=...)` runs the ported D-FINE
+  loop (AdamW param groups, EMA, AMP, grad clip, warmup + flat-cosine LR) with the same
+  **progress visualization as upstream**: a live console readout (`MetricLogger`) plus
+  TensorBoard scalars and a `loss_curve.png` under `output_dir` (needs
+  `pip install pydfine[train]`; W&B optional).
+
+- **COCO data + augmentation** — `dfine.train.dataset.build_coco_dataloader(img_folder,
+  ann_file, cfg=...)` gives a ready `(images, targets)` loader (contiguous-label remap,
+  multi-scale collate); pass `transforms=dfine.train.augment.train_transforms(imgsz,
+  stop_epoch=...)` for D-FINE's full augment pipeline (photometric distort, zoom-out,
+  IoU-crop, H-flip) with the two-phase no-aug tail. Feeds straight into `DFINE.train`.
+
+- **Have a YOLO dataset?** Convert it once — `dfine convert path/to/yolo path/to/coco`
+  (or `dfine.yolo_to_coco(...)`) — then `DFINE.train(data="path/to/coco")`. It reads the
+  `images/<split>` + `labels/<split>` layout (and `data.yaml` names) and writes the COCO
+  layout with 0-indexed categories that line up with the model's labels.
+
+Only `export` (Phase 3) remains. See [`docs/ROADMAP.md`](docs/ROADMAP.md) for the full
+status.
+
+```python
+from dfine import DFINEConfig
+
+cfg = DFINEConfig.preset("l", num_classes=3)   # verified upstream defaults
+cfg = DFINEConfig.preset("n")                  # 2-level, hidden_dim=128
+```
+
+The ported modules already run end-to-end (needs the `torch` extra installed):
+
+```python
+import torch
+from dfine import DFINEConfig
+from dfine.backends.native import HGNetv2, HybridEncoder, DFINETransformer
+
+cfg = DFINEConfig.preset("l", num_classes=80)
+backbone = HGNetv2.from_config(cfg).eval()
+encoder = HybridEncoder.from_config(cfg).eval()
+decoder = DFINETransformer.from_config(cfg).eval()
+
+out = decoder(encoder(backbone(torch.randn(1, 3, cfg.imgsz, cfg.imgsz))))
+# out["pred_logits"]: (1, 300, 80)   out["pred_boxes"]: (1, 300, 4)  [cxcywh, 0..1]
+```
+
+> The one-class `DFINE(...)` façade at the top of this README works today for
+> **inference** (`predict`/`load`/`from_pretrained`), **training** — both
+> `train(data="coco/", epochs=...)` (a standard COCO root; the loaders are built for
+> you) and `train(train_loader, epochs=...)` (a hand-built loader) — and **COCO
+> evaluation** (`val(data="coco/")` → the 12 named COCO metrics, also run each epoch
+> during `train`), all with the `pydfine[train]` extra. **Multi-GPU** is a single kwarg:
+> `train(data="coco/", devices=N)` spawns one DDP worker per GPU (or launch with
+> `torchrun` and call `train(...)` as usual). `export` is phase-stubbed and raises a
+> clear "arriving in Phase N" until implemented.
+
+## Why this exists
+
+Upstream D-FINE is an excellent research repo, but using it means editing YAML,
+copying config include-trees, and launching scripts. This library turns all of that
+into one importable, fully-typed class with presets — so a developer can go from
+`pip install` to a trained custom detector without touching a config file.
+
+## For contributors and AI agents
+
+This project is built to be developed largely by coding agents (Claude Code / any
+agent that reads `AGENTS.md`). Start here:
+
+| File | Purpose |
+|---|---|
+| [`AGENTS.md`](AGENTS.md) | **Canonical agent guide** — architecture, conventions, workflow, commands, definition of done. Read first. |
+| [`CLAUDE.md`](CLAUDE.md) | Claude Code–specific notes; defers to `AGENTS.md`. |
+| [`docs/ARCHITECTURE.md`](docs/ARCHITECTURE.md) | How D-FINE works and how we re-shape it into Python. |
+| [`docs/CONFIG_REFERENCE.md`](docs/CONFIG_REFERENCE.md) | Every model parameter, default, and per-size preset. The heart of the "one class, many params" design. |
+| [`docs/ROADMAP.md`](docs/ROADMAP.md) | Phased, checkbox task plan. |
+
+## License & attribution
+
+pydfine is licensed under the [Apache License 2.0](LICENSE).
+
+It is a **derivative work of D-FINE**
+([Peterande/D-FINE](https://github.com/Peterande/D-FINE), Apache-2.0, © 2024 The
+D-FINE Authors): the model is a native port of upstream `src/`, with layer and
+parameter names preserved so released `.pth` checkpoints load unchanged. Every ported
+module under `dfine/backends/native/` carries a per-file header crediting its source
+and describing the changes. See [`NOTICE`](NOTICE) for the full attribution, including
+D-FINE's own lineage (RT-DETR, DETR, PaddleDetection).

pydfine-0.0.1/dfine/__init__.py

@@ -0,0 +1,48 @@
+"""dfine — a config-first, ultralytics-style wrapper around D-FINE.
+
+The whole model is configured by typed params on one class. Phase 0/1 ships the
+config surface (``DFINEConfig``, presets); ``DFINE`` and ``Results`` land with the
+inference backend (Phase 2) and are exposed lazily so importing this package never
+requires torch until you actually build a model.
+"""
+
+from __future__ import annotations
+
+from typing import Any
+
+from .config import SIZE_PRESETS, SIZES, DFINEConfig, list_presets
+from .convert import yolo_to_coco
+from .registry import list_checkpoints
+
+__version__ = "0.0.1"
+
+__all__ = [
+    "DFINEConfig",
+    "SIZE_PRESETS",
+    "SIZES",
+    "list_presets",
+    "list_checkpoints",
+    "yolo_to_coco",
+    "__version__",
+]
+
+# Inference symbols live in modules that import torch; expose them lazily so a bare
+# `import dfine` (config/CLI only) never requires the torch extra.
+_LAZY = {
+    "DFINE": "model",
+    "Results": "results",
+    "Boxes": "results",
+}
+
+
+def __getattr__(name: str) -> Any:
+    if name in _LAZY:
+        try:
+            mod = __import__(f"dfine.{_LAZY[name]}", fromlist=[name])
+        except ImportError as exc:  # torch/torchvision/pillow not installed
+            raise AttributeError(
+                f"dfine.{name} needs the inference deps — install with "
+                f"`pip install pydfine[torch]` (missing: {exc.name})."
+            ) from exc
+        return getattr(mod, name)
+    raise AttributeError(f"module 'dfine' has no attribute {name!r}")