face-occlusion-classification (ONNX export)

ONNX export of the ConvNeXt-Small face-occlusion classifier from LamKser/face-occlusion-classification (MIT). All credit for the model and training goes to the original author; this repo only re-packages a trained checkpoint as ONNX for inference.

Files

face_occlusion.onnx — exported with their convert_2_onnx.py (opset 12).

I/O

Input: [1, 3, 224, 224] float32, RGB, scaled to [0,1] then ImageNet-normalized (mean [0.485,0.456,0.406], std [0.229,0.224,0.225]), NCHW.
Output: [1, 2] logits — softmax index 1 = occluded.

License & attribution

Released under the MIT License, inherited from the upstream project. See the original repository for the full copyright notice and training details.

Downloads last month: -; Downloads are not tracked for this model. How to track