feat: For QDQ models, disable constant folding and ensure the conv weights are transposed without using an explicit op#1856

fatcat-z merged 1 commit intoonnx:masteronnx/tensorflow-onnx:masterfrom

peri044:qdq_finalperi044/tensorflow-onnx:qdq_finalCopy head branch name to clipboard