An error (error message) occurred during the usage of PatchTSTEstimator.

[WeiBrian1994]

`import pandas as pd
import matplotlib.pyplot as plt

from gluonts.dataset.pandas import PandasDataset
from gluonts.dataset.split import split
from gluonts.torch.model.patch_tst import PatchTSTEstimator

Load data from a CSV file into a PandasDataset

df = pd.read_csv(
"data/ETTh1.csv",
index_col=0,
parse_dates=True,
)
dataset = PandasDataset(df, target="OT")

Split the data for training and testing

training_data, test_gen = split(dataset, offset=-36)
test_data = test_gen.generate_instances(prediction_length=12, windows=3)

Train the model and make predictions

model = PatchTSTEstimator(
prediction_length=12, patch_len = 96, trainer_kwargs={"max_epochs": 5}
).train(training_data)’

When running this code, it encounters a RuntimeError stating that mat1 and mat2 must have the same dtype, but got Double and Float. Although your dataset is indeed float64, you expect that PandasDataset would convert it to float32 automatically. However, the error indicates that there's a mismatch in the data types somewhere in the computation process.

[Hafsa-ds]

Hi,
I am receiving the same error with TFT. Did you manage to solve it ?