# SEA-RAFT
SEA-RAFT (Simple, Efficient, Accurate RAFT) is a RAFT descendant by Wang, Lipson and Deng (ECCV 2024). It predicts dense optical flow between two frames and, unlike the other flow families in LibreYOLO, also a per-pixel confidence map. LibreYOLO ports the upstream inference code and mirrors its S and M checkpoints.
Tasks: Optical flow. Install: pip install libreyolo.
Verified against LibreYOLO v1.8.0.

SEA-RAFT keeps RAFT's all-pairs correlation pyramid, regresses an
initial flow directly, refines it with a ConvNeXt update block and is trained
with a mixture-of-Laplace loss. The predicted mixture variance is what
LibreYOLO turns into `flow.confidence`. It is the family the
[optical flow task page](/docs/tasks/optical-flow) uses in its examples; pick it
when you want a per-pixel confidence next to the flow.

## Install

SEA-RAFT needs no optional extra. Everything it imports is in the base install.

```bash
pip install libreyolo
```

## Predict

Weights download from the LibreYOLO Hugging Face organization on first use.
The class `LibreSEARAFT` is also exported from `libreyolo` if you prefer to build the
model directly.

**Python**

```python
from libreyolo import LibreYOLO

model = LibreYOLO("LibreSEARAFTm-flow.pt")   # weights download on first use
results = model(["frame_0001.jpg", "frame_0002.jpg"])

flow = results[0].flow
print(flow.data.shape)        # (H, W, 2) float32, (dx, dy) in pixels
print(flow.mean_magnitude)    # mean displacement in pixels
```

**Video**

```python
from libreyolo import LibreYOLO

model = LibreYOLO("LibreSEARAFTm-flow.pt")

# N frames give N - 1 results; result i is the flow from frame i to i + 1.
for result in model("video.mp4", stream=True, save=True, save_flow=True):
    print(result.flow.max_magnitude)
```

**CLI**

```bash
libreyolo predict model=LibreSEARAFTm-flow.pt source=frame_0001.jpg,frame_0002.jpg save=true
libreyolo predict model=LibreSEARAFTm-flow.pt source=video.mp4 save=true arrows=true save_flow=true
```

Flow needs at least two frames: a single image raises. A list of images, a
directory sorted by name, a video or a stream is one sequence, and `N` frames
give `N - 1` results. `result.flow` is a `FlowField` with an `(H, W, 2)` float32
array on the original canvas of the first frame; `flow.data[y, x]` is
`(dx, dy)` in pixels, `+x` right and `+y` down. `flow.confidence` is an `(H, W)` map in `[0, 1]` derived from the
predicted mixture variance: higher means the network is more certain of the
vector at that pixel. `result.boxes` is
`None`. `save=True` writes the Middlebury colour wheel, `arrows=True` or
`warp=True` pick the other views, and `save_flow=True` also writes each pair's
raw field as a `.flo` file. The [optical flow task](/docs/tasks/optical-flow)
page describes the payload and views in full.

## Variants

| Size | Encoders | Refinement iterations | File |
| --- | --- | ---: | --- |
| `s` | ResNet18 | 4 | `LibreSEARAFTs-flow.pt` |
| `m` | ResNet34 | 4 | `LibreSEARAFTm-flow.pt` |
| `l` | ResNet34, the `m` weights | 12 | `LibreSEARAFTl-flow.pt` |

Upstream ships no separate L weights: `l` is the `m` checkpoint run with 12
iterations instead of 4. All three are upstream's
Tartan-C-T-TSKH-spring540x960 checkpoints. `LibreSEARAFT` defaults to size
`m`. Frames up to 960 px on the long side, the training crop width, run at
native resolution padded to a multiple of 8; larger frames are downscaled and
the field is scaled back. `imgsz` changes that cap.

Training is not offered: `train()` raises `NotImplementedError`.

## Validate

`val()` runs the shared flow validator on a flow dataset (frames per sequence
folder, Middlebury `.flo`, KITTI 16-bit PNG or `.npy` ground truth; see the
[dataset format](/docs/tasks/optical-flow#dataset-format)). It reports the mean
end-point error `metrics/epe`, `metrics/px1` and `metrics/px3` (share of valid
pixels under 1 and 3 px of error) and the KITTI outlier rate
`metrics/fl_all`.

**Python**

```python
from libreyolo import LibreYOLO

model = LibreYOLO("LibreSEARAFTm-flow.pt")
metrics = model.val(data="my_flow.yaml")

print(metrics["metrics/epe"])      # mean end-point error, pixels
print(metrics["metrics/px1"])      # share of pixels under 1 px error
print(metrics["metrics/fl_all"])   # KITTI outlier rate
```

**CLI**

```bash
libreyolo val model=LibreSEARAFTm-flow.pt data=my_flow.yaml
```

## Export

| Task | ONNX | TorchScript | ExecuTorch | TensorRT | OpenVINO | Paddle | MNN | RKNN | Axelera | ncnn | TFLite | CoreML | Core AI |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
| Optical flow |  |  |  |  |  |  |  |  |  |  |  |  |  |

A "yes" means the export is supported. An empty cell means the exporter refuses that combination.

Export is not available yet for any format. Optical flow export is blocked
until the two-input graph (`image1`, `image2`) and the backend vector-rescale
contract of the flow task are implemented for a family.

## Checkpoints

Every published weight file for this family.

| File | Input (px) | Task | Weights license |
| --- | --- | --- | --- |
| `LibreSEARAFTs-flow.pt` |  | Optical flow | bsd-3-clause |
| `LibreSEARAFTm-flow.pt` |  | Optical flow | bsd-3-clause |
| `LibreSEARAFTl-flow.pt` |  | Optical flow | bsd-3-clause |

## Licensing

Check the license on the Hugging Face repository of the specific weights you download. That repository is authoritative and licenses are not always uniform across a family. This is a description of the licenses involved, not legal advice.

- Original work: SEA-RAFT, Princeton Vision & Learning Lab
- Upstream license: BSD-3-Clause
- Upstream source: https://github.com/princeton-vl/SEA-RAFT
- LibreYOLO code: MIT
- Weights: BSD-3-Clause, republished at https://huggingface.co/LibreYOLO
- Interpretation: BSD-3-Clause is a permissive license: the code and weights can be used in commercial and closed-source products. Redistributions must keep the copyright notice, the license conditions and the disclaimer, and the authors' names may not be used to endorse derived products without permission. An inference subset of the upstream core/ package at commit 9137517b is vendored under libreyolo/models/searaft/_vendor/ with its BSD-3-Clause LICENSE; LibreYOLO's wrapper code is MIT. The weights are the upstream Tartan-C-T-TSKH-spring540x960 S and M checkpoints (MemorySlices on Hugging Face), which upstream publishes without a separate weight license; LibreYOLO mirrors them under the repository's BSD-3-Clause license and states that on the cards. Size l reuses the m weights with 12 refinement iterations instead of 4. No gating; inference only.

SEA-RAFT is the work of Wang, Lipson and Deng at the
[Princeton Vision & Learning Lab](https://github.com/princeton-vl/SEA-RAFT).
An inference subset of the upstream code is vendored under its BSD-3-Clause
licence. Upstream publishes the S and M weights without a separate weight
licence; LibreYOLO mirrors them under the repository's BSD-3-Clause licence,
without gating. BSD-3-Clause allows commercial and closed-source use as long
as the copyright notice, conditions and disclaimer are kept.

## Citation

```bibtex
@article{wang2024sea,
  title={SEA-RAFT: Simple, Efficient, Accurate RAFT for Optical Flow},
  author={Wang, Yihan and Lipson, Lahav and Deng, Jia},
  journal={arXiv preprint arXiv:2405.14793},
  year={2024}
}
```

Copied from https://github.com/princeton-vl/SEA-RAFT
