plot_prediction throw ValueError while using Exogenous Infusion

[omriargaman]

When running plot_predictions with Exogenous Infusion (as provided on example notebook) i get an error message:

ValueError Traceback (most recent call last)
in <cell line: 1>()
----> 1 plot_predictions(
2 model=finetune_forecast_trainer.model,
3 dset=test_dataset,
4
5 plot_dir=os.path.join(OUT_DIR, "bike_sharing"),

9 frames
/usr/local/lib/python3.10/dist-packages/tsfm_public/models/tinytimemixer/modeling_tinytimemixer.py in forward(self, base_forecasts, past_values, future_values)
696
697 if self.exogenous_channel_count > 0 and future_values is None:
--> 698 raise ValueError("future_values cannot be none when we have exogenous channels.")
699
700 if self.exogenous_channel_count > 0:

ValueError: future_values cannot be none when we have exogenous channels.

although the dataset test_dataset holds future_values

[wgifford]

@omriargaman Thank you for the bug report. We are investigating and will report back.

[wgifford]

Please try the latest from main. This will be included in the next release.

[wgifford]

Closing for now, please reopen if there is an issue.