SKLearn regression model with multiple outputs

## 🌱 Describe your Feature Request
I want to convert a `sklearn.linear_model.LinearRegression` model with multiple output targets (149 to be exact) but it seems like the [converter](https://apple.github.io/coremltools/source/coremltools.converters.sklearn.html#module-coremltools.converters.sklearn._converter) only supports a single output. This looks like an artificial limitation, is it possible to support multiple outputs?

## Describe alternatives you've considered
A multifunction-model is possible but quite annoying to work with. Since this is a simple linear function I think it can be first converted to neural network and then converted that way but this won't work as easily for other models.

Thanks!

Start with the scikit-learn converter module linked in the issue and trace how LinearRegression models are handled. Determine how the converter currently limits outputs, then validate support for 149 output targets and confirm the converted model preserves all outputs; the issue names no test file.

Written by the indexing model from the issue text.