imagesstatues

An audio event classifier trained on the AudioSet dataset to predict audio events from the AudioSet ontology.

yamnet

yolov5_boresv2

ml-yolo-v4

yolov5

road-sing-predictor

testmodel

Hand landmark prediction with MediaPipe Hands in Unity Engine.

## Importing the Predictor
Follow the setup instructions on [the predictor implementation](https://github.com/natml-hub/BlazePalm).

## Using the Predictor
```csharp
// Fetch model data from NatML
var modelData = await MLModelData.FromHub("@natml/blazepalm-landmark");
// Deserialize the model
var model = await MLEdgeModel.Create(modelData);
// Create the predictor
var predictor = new BlazePalmPredictor(model);
```



blazepalm-landmark

Palm detection from Google MediaPipe.

## Importing the Predictor
Follow the setup instructions on [the predictor implementation](https://github.com/natml-hub/BlazePalm-Detector).

## Using the Predictor
```csharp
// Fetch model data from NatML
var modelData = await MLModelData.FromHub("@natml/blazepalm-detector");
// Deserialize the model
var model = await MLEdgeModel.Create(modelData);
// Create the predictor
var predictor = new BlazePalmDetector(model);
```



blazepalm-detector

MediaPipe Selfie Segmentation for human matting.

## Importing the Predictor
Follow the setup instructions on [the predictor implementation](https://github.com/natml-hub/Meet).

## Using the Predictor
```csharp
// Fetch model data from NatML
var modelData = await MLModelData.FromHub("@natml/meet");
// Deserialize the model
var model = await MLEdgeModel.Create(modelData);
// Create the predictor
var predictor = new MeetPredictor(model);
```



meet

Pose landmark prediction from Google MediaPipe.

## Importing the Predictor
Follow the setup instructions on [the predictor implementation](https://github.com/natml-hub/BlazePose).

## Using the Predictor
```csharp
// Fetch model data from NatML
var modelData = await MLModelData.FromHub("@natml/blazepose-landmark");
// Deserialize the model
var model = await MLEdgeModel.Create(modelData);
// Create the predictor
var predictor = new BlazePosePredictor(model);
```



blazepose-landmark

Pose detection from Google MediaPipe.

## Importing the Predictor
Follow the setup instructions on [the predictor implementation](https://github.com/natml-hub/BlazePose-Detector).

## Using the Predictor
```csharp
// Fetch model data from NatML
var modelData = await MLModelData.FromHub("@natml/blazepose-detector");
// Deserialize the model
var model = await MLEdgeModel.Create(modelData);
// Create the predictor
var predictor = new BlazePoseDetector(model);
```



blazepose-detector

Multi-person pose detection from Google MediaPipe.

## Importing the Predictor
Follow the setup instructions on [the predictor implementation](https://github.com/natml-hub/MoveNet-Multipose).

## Using the Predictor
```csharp
// Fetch model data from NatML
var modelData = await MLModelData.FromHub("@natml/movenet-multipose");
// Deserialize the model
var model = await MLEdgeModel.Create(modelData);
// Create the predictor
var predictor = new MoveNetMultiposePredictor(model);
```



movenet-multipose

High performance object detector based on YOLO series.

## Importing the Predictor
Follow the setup instructions on [the predictor implementation](https://github.com/natmlx/yolox-unity).

## Using the Predictor
```csharp
// Fetch model data from NatML
var modelData = await MLModelData.FromHub("@natsuite/yolox");
// Deserialize the model
var model = await MLEdgeModel.Create(modelData);
// Create the predictor
var predictor = new YOLOXPredictor(model, modelData.labels);
```

## References
- [Model info](https://arxiv.org/abs/2107.08430)


yolox

nanodet

Realtime single-stage face detector.

## Importing the Predictor
Follow the setup instructions on [the predictor implementation](https://github.com/natml-hub/DBFace).

## Using the Predictor
```csharp
// Fetch model data from NatML
var modelData = await MLModelData.FromHub("@natsuite/dbface");
// Deserialize the model
var model = await MLEdgeModel.Create(modelData);
// Create the predictor
var predictor = new DBFacePredictor(model);
```

## References
- [Model info](https://github.com/dlunion/DBFace)


dbface

Ultra-high performance face detection architecture.

## Importing the Predictor
Follow the setup instructions on [the predictor implementation](https://github.com/natml-hub/YuNet).

## Using the Predictor
```csharp
// Fetch model data from NatML
var modelData = await MLModelData.FromHub("@natsuite/yunet");
// Deserialize the model
var model = await MLEdgeModel.Create(modelData);
// Create the predictor
var predictor = new YuNetPredictor(model);
```

## References
- [Model info](https://github.com/ShiqiYu/libfacedetection)


yunet

Lightweight face detection from Google MediaPipe.

## Importing the Predictor
Follow the setup instructions on [the predictor implementation](https://github.com/natml-hub/BlazeFace).

## Using the Predictor
```csharp
// Fetch model data from NatML
var modelData = await MLModelData.FromHub("@natsuite/blazeface");
// Deserialize the model
var model = await MLEdgeModel.Create(modelData);
// Create the predictor
var predictor = new BlazeFacePredictor(model);
```

## References
- [Model info](https://google.github.io/mediapipe/solutions/face_detection.html)


blazeface

CenterFace anchor-free face detection.

## Importing the Predictor
Follow the setup instructions on [the predictor implementation](https://github.com/natml-hub/CenterFace).

## Using the Predictor
```csharp
// Fetch model data from NatML
var modelData = await MLModelData.FromHub("@natsuite/centerface");
// Deserialize the model
var model = await MLEdgeModel.Create(modelData);
// Create the predictor
var predictor = new CenterFacePredictor(model);
```

## References
- [Model info](https://github.com/Star-Clouds/CenterFace)


centerface

Single Shot Detector Lite for general object detection.

## Importing the Predictor
Follow the setup instructions on [the predictor implementation](https://github.com/natml-hub/SSD-Lite).

## Using the Predictor
```csharp
// Fetch model data from NatML
var modelData = await MLModelData.FromHub("@natsuite/ssd-lite");
// Deserialize the model
var model = await MLEdgeModel.Create(modelData);
// Create the predictor
var predictor = new SSDLitePredictor(model, modelData.labels);
```

## References
- [Model info](https://arxiv.org/abs/1512.02325)


ssd-lite

Ultra-light fast generic face detector.

## Importing the Predictor
Follow the setup instructions on [the predictor implementation](https://github.com/natml-hub/UltraFace).

## Using the Predictor
```csharp
// Fetch model data from NatML
var modelData = await MLModelData.FromHub("@natsuite/ultraface");
// Deserialize the model
var model = await MLEdgeModel.Create(modelData);
// Create the predictor
var predictor = new UltraFacePredictor(model);
```

## References
- [Model info](https://github.com/Linzaer/Ultra-Light-Fast-Generic-Face-Detector-1MB)


ultraface

Robust video matting for alpha compositing.

## Importing the Predictor
Follow the setup instructions on [the predictor implementation](https://github.com/natml-hub/Robust-Video-Matting).

## Using the Predictor
```csharp
// Fetch model data from NatML
var modelData = await MLModelData.FromHub("@natsuite/robust-video-matting");
// Deserialize the model
var model = await MLEdgeModel.Create(modelData);
// Create the predictor
var predictor = new RobustVideoMattingPredictor(model);
```

## References
- [Model info](https://peterl1n.github.io/RobustVideoMatting)


robust-video-matting

DistilBERT language model for sentence filling.

## Importing the Predictor
Follow the setup instructions on [the predictor implementation](https://assetstore.unity.com/packages/tools/integration/distilbert-lm-autocomplete-with-ai-202022).

## Using the Predictor
```csharp
// Fetch model data from NatML
var modelData = await MLModelData.FromHub("@natsuite/distilbert-masked-lm");
// Deserialize the model
var model = await MLEdgeModel.Create(modelData);
// Create the predictor
var predictor = new DistilBERTMaskedLM(model);
```

## References
- [Model info](https://huggingface.co/transformers/model_doc/distilbert.html#distilbertformaskedlm)


distilbert-masked-lm

DistilBERT language model for question answering.

## Importing the Predictor
Follow the setup instructions on [the predictor implementation](https://assetstore.unity.com/packages/tools/integration/distilbert-qa-question-answering-with-ai-201527).

## Using the Predictor
```csharp
// Fetch model data from NatML
var modelData = await MLModelData.FromHub("@natsuite/distilbert-qa");
// Deserialize the model
var model = await MLEdgeModel.Create(modelData);
// Create the predictor
var predictor = new DistilBERTQA(model, modelData.labels);
```

## References
- [Model info](https://huggingface.co/transformers/model_doc/distilbert.html#distilbertforquestionanswering)


distilbert-qa

Hair matte segmentation.

## Importing the Predictor
Follow the setup instructions on [the predictor implementation](https://github.com/natml-hub/Hair-Matte-Net).

## Using the Predictor
```csharp
// Fetch model data from NatML
var modelData = await MLModelData.FromHub("@natsuite/hair-matte-net");
// Deserialize the model
var model = await MLEdgeModel.Create(modelData);
// Create the predictor
var predictor = new HairMatteNetPredictor(model);
```



hair-matte-net

Realtime pose detection in 3D with augmented reality.

## Importing the Predictor
Follow the setup instructions on [the predictor implementation](https://github.com/natml-hub/MoveNet-3D).

## Using the Predictor
```csharp
// Fetch model data from NatML
var modelData = await MLModelData.FromHub("@natsuite/movenet-3d");
// Deserialize the model
var model = await MLEdgeModel.Create(modelData);
// Create the predictor
var predictor = new MoveNet3DPredictor(model);
```

## References
- [Model info](https://blog.tensorflow.org/2021/05/next-generation-pose-detection-with-movenet-and-tensorflowjs.html)


movenet-3d

Hand pose detection from MediaPipe.

## Importing the Predictor
Follow the setup instructions on [the predictor implementation](https://github.com/natml-hub/Hand-Pose).

## Using the Predictor
```csharp
// Fetch model data from NatML
var modelData = await MLModelData.FromHub("@natsuite/hand-pose");
// Deserialize the model
var model = await MLEdgeModel.Create(modelData);
// Create the predictor
var predictor = new HandPosePredictor(model);
```

## References
- [Model info](https://google.github.io/mediapipe/solutions/hands.html)


hand-pose

Lightweight object classification model optimized for mobile applications.

## Importing the Predictor
Follow the setup instructions on [the predictor implementation](https://github.com/natml-hub/MobileNet-v2).

## Using the Predictor
```csharp
// Fetch model data from NatML
var modelData = await MLModelData.FromHub("@natsuite/mobilenet-v2");
// Deserialize the model
var model = await MLEdgeModel.Create(modelData);
// Create the predictor
var predictor = new MobileNetv2Predictor(model, modelData.labels);
```



mobilenet-v2

MoveNet Lightning architecture for single body pose detection.

## Importing the Predictor
Follow the setup instructions on [the predictor implementation](https://github.com/natml-hub/MoveNet).

## Using the Predictor
```csharp
// Fetch model data from NatML
var modelData = await MLModelData.FromHub("@natsuite/movenet");
// Deserialize the model
var model = await MLEdgeModel.Create(modelData);
// Create the predictor
var predictor = new MoveNetPredictor(model);
```

## References
- [Model info](https://blog.tensorflow.org/2021/05/next-generation-pose-detection-with-movenet-and-tensorflowjs.html)
