[BUG] DLinear fails to predict >1 variates in presence of static covariates

[felixdivo]

Describe the bug
Predicting two or more variates fails if there are static covariates present and shall be used.

To Reproduce

import pandas as pd
from darts.datasets import ETTh1Dataset
from darts.models import DLinearModel

series = ETTh1Dataset().load().astype("float32")

target = series["HUFL"].stack(series["LULL"]).with_static_covariates(pd.Series([3, 4]))
past_cov = series["MULL"]

model = DLinearModel(16, 4, use_static_covariates=True)
model.fit(series=target, past_covariates=past_cov)

---> RuntimeError: mat1 and mat2 shapes cannot be multiplied (32x2 and 4x8)

Expected behavior
Not a crash, but nice forecasts.

System (please complete the following information):

• Python version: 3.10.13
• PyTorch version: 2.1.0
• darts version: 0.26.0

Additional context
Possibly related to #1873 and #2035.