Simplify the structure of val_set

[WenjieDu]

PyPOTS/pypots/imputation/saits/model.py

Lines 268 to 275 in d457629

	# check if X_intact contains missing values
	if np.isnan(val_set["X_intact"]).any():
	val_set["X_intact"] = np.nan_to_num(val_set["X_intact"], nan=0)
	logger.warning(
	"X_intact shouldn't contain missing data but has NaN values. "
	"PyPOTS has imputed them with zeros by default to start the training for now. "
	"Please double-check your data if you have concerns over this operation."
	)

For imputation models, we expect X to contain missing values, while for X_intact, we need it with missing data filled with 0. We also need the additional indicating_mask recording positions for error calculation, but this is not necessary. Because indicating_mask can easily be calculated like torch.isnan(X)^torch.isnan(X_intact) if missing values are left as NaN in X and X_intact.