[ENH] Adding Support of Distribution Loss in v2 Models.

[Muhammad-Rebaal]

Is your feature request related to a problem? Please describe.
Currently there isn't any support of Distribution Loss in v2 models

Describe the solution you'd like
The V2 BaseModel _base_model_v2.py should detect when the passed loss is a DistributionLoss from metrics/distributions.py and automatically infer output_size from len(loss.distribution_arguments), call loss.rescale_parameters() in the training loop, and ensure target_scale is passed through from the V2 DataModule batch the full DistributionLoss contract map_x_to_distribution, to_quantiles, sample is already implemented and only needs to be wired into the V2 model layer.