openvino

9131e4e7 - [ONNX_FE] QuantizeLinear: create zero rank shapes in case of per tensor quantization (#33822)

[ONNX_FE] QuantizeLinear: create zero rank shapes in case of per tensor quantization (#33822) ### Details: - *Make `is_per_tensor_quantization` check more general* - *Avoid zero point / FQ ranges constants with non zero rank creation in case of per tensor quantization so that when the data input of the `FakeQuantize` has an empty shape (rank=0), the corresponding FQ constants do not end up having a higher rank (e.g., rank‑1). This ensures that the shapes of the range constants remain consistent with the shape of the data input in per‑tensor quantization scenarios.* ### Tickets: - *N\A*