TerminatedWorkerError when fitting the REaLTabFormer model

[wennaq]

Hi there,

I'd like to train a REalTabFormer model for regular data. I ran the code snippet below, and the code crashed at rtf_model.fit(df) after running for 3 hours with the error message shown in the screeshot.

Could you please provide some insights into what could have triggered the error and how I can resolve it? Thanks!

Code Snippet

import pandas as pd
from realtabformer import REaLTabFormer

df = pd.read_csv("foo.csv")

# NOTE: Remove any unique identifiers in the
# data that you don't want to be modeled.

# Non-relational or parent table.
rtf_model = REaLTabFormer(
    model_type="tabular",
    gradient_accumulation_steps=4,
    logging_steps=100)

# Fit the model on the dataset.
# Additional parameters can be
# passed to the `.fit` method.
rtf_model.fit(df)

Error Message

Environment

• realtabformer 0.2.2
• pandas 2.2.2
• python 3.9
• platform: Jupyter Notebooks
• one GPU