This change enables more customization with operand quantization, and
generalizes the patterns QuantizeOperands and QuantizeTransposeOperands
to QuantizeOperandsPastCommutingOps.
This allows for passing quantization through operations which are
functionally unaffected by quantization, such as view-like ops. The
purpose of this change is to address a myriad of quantization issues
seen in quantized onnx models that have some reshape-like operations
sandwiched in between a dequant and something like a matmul (whose other
operand is immediately quantizable).
Leaning on the QDQ functionality in torch we can support the QLinearConv
operation by piggybacking through `torch.Convolution`. This includes
some changes such as allowing the `onnx` rewriter to run recursively.
Doing so allows `QLinearConv` to decopmose to `onnx.Convolution` which
is then lowered to `torch`.
Linalg has quantized specific operations. We can lower to these
operations when there is a known zeropoint and scale operations. This
allows the `convolution` to occur with lower bitwidth's, improving the
overall performance.
This includes custom op matching for decomposed operations and fusing
dequantization into dense operations. As a validation we compare
to the dequant+mm torch implementation.
zjgarvey
Rob Suderman