[FP8] update and add necessary test for the ONNX Ops for FP8 dtypes

[umangyadav]

At the time of writing this issue, ONNX has FP8 support for the following ops. ONNX will keep adding FP8 support for more operators.

• castlike
• cast
• constant
• constantofshape
• dequantizelinear
• identity
• If
• isinf
• isnan
• loop
• QLinearMatmul
• QuantizeLinear
• reshape
• Scan
• Shape
• Size

As part of this issue, need to figure out if for each of the ops if anything requires an update in onnx parsing logic and/or testing.

In most cases, FP8 dtype would be handled transparently and won't require any change in parsing but may need to add tests for Fp8 in onnx_test.cpp.

Things to consider when updating ops is "saturation" and "rounding" mode for FP8.