Design UX for model quantization

[leoll2]

Create a design for the model quantization feature according to the requirement #4944

Differences w.r.t. Geti 2.x:

• the quantization job exposes a couple of configurable input parameters (max accuracy drop, max calibration dataset size)
• the user can quantize the same base model (OV FP16) multiple times, possibly with different configurations
• since there is no jobs panel, the progress of the quantization job could be displayed in the "Models" page similarly to how it's done for the training job