gradient optimizer for PyTorch Temporal Fusion Transformer (TFT) #17036
-
Dear Community, I am using the PyTorch TemporalFusionTransformer implementation along with the PyTorch lightning Trainer. (https://pytorch-forecasting.readthedocs.io/en/latest/api/pytorch_forecasting.models.temporal_fusion_transformer.TemporalFusionTransformer.html + https://pytorch-lightning.readthedocs.io/en/stable/common/trainer.html) Unfortuantely, i can't find any information about the default learning rate optimizer is used. Does anyone have any idea? As far as I understand it correctly, the authors of TFT use the Adam optimizer with learning_rate and max_gradient_norm: Best regards Stefan |
Beta Was this translation helpful? Give feedback.
Replies: 1 comment
-
I finaly found it :) |
Beta Was this translation helpful? Give feedback.
I finaly found it :)
Pytorch uses the Ranger optimizer as default.