Why does my XGB multiclass classification problem get a multioutput error?

Question

I'm trying to learn XGB with a multiclass classification problem (Kaggle's Otto Group Product Classification). I keep getting the following error

XGBoostError: [04:49:48] ../include/xgboost/objective.h:98: multioutput is not supported by current objective function

Here's my code:

import xgboost as xgb

params = {
    'objective': 'multi:softmax',
    'num_class': 9,
    'max_depth': 5,
    'learning_rate': 0.1,
    'subsample': 0.8,
    'colsample_bytree': 0.8,
    'seed': 42
}

dtrain = xgb.DMatrix(X_train, label=y_train)

num_rounds = 100
xgb_model = xgb.train(params, dtrain, num_rounds)

X_train is a 50,000 x 93 dataframe and y_train is a 50,000 x 9 numpy area. Additionally, y_train is a one-hot encoded array representing the 9 categories.

I'm not sure why I'm getting a multi-ouput error since this isn't a multi-output problem.