I want to load Parquet row groups into batches. below code working fine on my local system-

# dbfs:/output/scaled.parquet
batch_size=2
with make_batch_reader('file:/dbfs:/output/scaled.parquet', num_epochs=1,shuffle_row_groups=False) as train_reader:
    train_ds = make_petastorm_dataset(train_reader).unbatch().map(lambda x: (tf.convert_to_tensor(x))).batch(batch_size)
    for batch in train_ds:
        X_train = tf.reshape(batch,(2,1,15))
        model.fit(X_train,X_train)

But when i am trying on DataBricks it return given error TypeError: init() missing 2 required positional arguments: 'instance' and 'token'

• So i give the instance and token in make_batch_reader but know it's given TypeError: make_batch_reader() got an unexpected keyword argument 'instance'

Help to resolve the issue