bumie

Hi, Thank you for the quick response. Here's the full code def build_model(observed_time_series, data): smooth_seasonal = sts.SmoothSeasonal( period=365.25, # Annual period observed_time_series= observed_time_series, frequency_multipliers=[1, 2, 3], name='smooth_seasonal' ) autoregressive =...

Oh, I see. Perhaps the error is with Raytune then. Thank you!

Hi, It seems the error happens when I try to use distributed training with tfp. I replaced RayTune with Optuna, ran the model on a Dask Cluster with 2 GPUs...

Yes. Stack trace with RayTune and Optuna as the optimizer algorithm I updated the notebook you sent me with the code. https://colab.research.google.com/gist/bumie-e/4af2d03e9c446d2de374ebb03ccd41dd/untitled47.ipynb /usr/local/lib/python3.9/dist-packages/ray/tune/search/optuna/optuna_search.py:685: FutureWarning: IntUniformDistribution has been deprecated in v3.0.0....