Can we include some quantized models w.r.t. Quantization-aware Training?

Model Request

Which model would you like to see in the model zoo?

A quantized MobileNet (doesn't matter which version) could be fine. TensorFlow has published end to end quantized MobileNetV1.

Describe why this model is relevant

I am interested in the quantization representation of ONNX, and how it peserves quantization parameters which are generated by frameworks during quantization-aware training. I have read ONNX documents and operator spec, but a quantized model could have great impact to help people understand ONNX quantization. It won't be limited to MobileNet I think.

Model information

Source repository (e.g. Github link): TFLite quantized MobileNetV1 doc.
Model Zoo category (e.g. Vision): Vision or any.
Model Zoo sub-category (e.g. Image Classification): Image Classification or any.

Please feel free to ping me if any addition information is needed.

No repository file or test is named. Start by reviewing the model-zoo contribution structure and the linked TensorFlow quantized MobileNetV1 source, then determine how an ONNX model should preserve quantization-aware-training parameters. Done means a validated quantized MobileNet model is added to an appropriate vision classification category.