Which is the PyTorch equivalent Dynamic Quantization in ONNX? I have an ONNX model generated from a HuffingFace transformer.
