[quantization] Propagate qparam for expand

[stamalakhov]

This PR propagates qparam forward for torch.ops.aten.expand.default .

It occurs that q_i @ k_i.transpose(-2, -1) expands k_i to match q_i shape. The expand operation is not quantized, so that its output is float32. Although o1 optimizer from ONE from handle such cases, let's try to support it here.

To find unquantized nodes just look for expand_1 in Netron and check its output.

./ccex test -k quantization.pass.test_propagate_quant_param.ExpandTest

RUN unit tests with -k quantization.pass.test_propagate_quant_param.ExpandTest ...
test_s16 (quantization.pass.test_propagate_quant_param.ExpandTest) ... ok

----------------------------------------------------------------------
Ran 1 test in 0.011s

OK

Draft: #436
TICO-DCO-1.0-Signed-off-by: s.malakhov s.malakhov@partner.samsung.com

[mhs4670go]

LGTM