Converting Scikit-learn MLPRegressor to Tensorflow Keras model

Question

Scikit-learn's MLPRegressor class is as simple to implement as any other sklearn model. Here's what I've been using so far:

from sklearn.neural_network import MLPRegressor
from sklearn.pipeline import Pipeline
from sklearn.compose import TransformedTargetRegressor
from sklearn.preprocessing import StandardScaler

base = MLPRegressor(max_iter=50, hidden_layer_sizes=(100,), early_stopping=True, learning_rate="adaptive")
pipeline = Pipeline([('scaler', StandardScaler()), ('model', base)])
model = TransformedTargetRegressor(regressor=pipeline, transformer=StandardScaler())

What I'd like to do is implement something virtually identical in Tensorflow as a stepping stone to making a more complicated model with separate LSTM and Dense channels. Here's what I have so far:

from tensorflow.keras.layers import Input, Dense
from keras.callbacks import EarlyStopping
from tensorflow.keras.wrappers.scikit_learn import KerasRegressor

def simple_tf_model():
    dense_input = Input(shape=(rf_train.shape[1],))
    dense = Dense(100, activation="relu")(dense_input)
    dense = Dense(1)(dense)
    tf_model = Model(inputs=[dense_input], outputs=dense)
    tf_model.compile(loss="mse",optimizer="adam")
    return tf_model
     
# monitor='val_loss' with validation_split=0.1 based on the early_stopping and validation_fraction parameters of MLPRegressor
es = EarlyStopping(monitor='val_loss', mode='min', verbose=1)
# batch_size=200 used as equivalent to batch_size="auto" for MLPRegressor
tf_model = KerasRegressor(build_fn=simple_tf_model, batch_size=200, epochs=50, validation_split=0.1, callbacks=[es])
pipeline = Pipeline([('scaler', StandardScaler()), ('model', tf_model)])
model = TransformedTargetRegressor(regressor=pipeline, transformer=StandardScaler())
model.fit(rf_train, rf_train_y)

However, when I run this, I get noticeably better performance from the MLPRegressor model than from the Tensorflow version.

Am I doing something wrong in the Tensorflow implementation?