Tabular Model with 2 outputs -> one categorial and one regression

I would like to create a tabular model with multiple outputs. Currently, I can create a model with 2 regression outputs using the following code:

to = TabularPandas(
    procs=[Categorify, Normalize],
    cat_names=["cat1", 'cat2', 'cat3'],
    cont_names=['cont1', 'cont2', "cont3"],
    y_names=["y_reg", 'y_reg2'],
    y_block=RegressionBlock(n_out = 2),

After creating the model, I modify my losses and metrics using the following code:

import numpy as np

def board_loss(inp, true):
    return F.mse_loss(inp[:, 0], true[:, 0])

def exit_loss(inp, true): 
    return F.mse_loss(inp[:, 1], true[:, 1])

def combine_loss(inp, true): 
    return board_loss(inp,true)+exit_loss(inp,true)

def board_error(inp, true): 
    return rmse(inp[:, 0].exp(), true[:, 0].exp())

def exit_error(inp, true):
    return rmse(inp[:, 1].exp(), true[:, 1].exp())

def board_r2(inp, true):
    return np.corrcoef(inp[:, 0].exp().cpu().numpy(), true[:, 0].exp().cpu().numpy())[0][1]

def exit_r2(inp, true):
    return np.corrcoef(inp[:, 1].exp().cpu().numpy(), true[:, 1].exp().cpu().numpy())[0][1]

err_metrics = (board_error,exit_error, board_r2, exit_r2)
all_metrics = err_metrics+(board_loss,exit_loss)

I would now like to modify y2 to be a category while keeping y1 as a regression output. However, y_block in TabularPandas only allows for one type of output. Additionally, if I can achieve this, I would like to run y2 through a sigmoid and perhaps add additional layers for classification.