pytorch-forecasting

pytorch-forecasting copied to clipboard

[ENH] `TemporalFusionTransformer` - allow mixed precision training

2

### Description This PR modifies the attention mask in the TFT model from 1e-9 to float("inf") to allow Pytorch mixed precision training. Closes #1325, closes #285

Marcrb2

I have a general question about how to handle training where some of the covariates are themselves forecasts. I am building a TFT model where the decoder part of each...

arizzuto

Is the DeepVAR implementation correct?

2

I was reading the paper that the DeepVAR implementation is supposed to be based on (https://arxiv.org/abs/1910.03002) and I'm having trouble finding the Gaussian copula processes anywhere in the source code....

chododom

Fixed typo "monotone_constaints"

1

### Description monotone_constaints -> monotone_constraints ### Checklist - [ ] Linked issues (if existing) - [ ] Amended changelog for large changes (and added myself there as contributor) - [...

Luke-Chesley

Issue while installing pytorch-forecasting through pip

5

- PyTorch-Forecasting version: Can't Install due to wheel issue. - PyTorch version: Can't install same wheel issue - Python version: 3.11.5 - Operating System: Windows Issue faced while installing pytorch-forecasting...

DS-mehul

Fixed RuntimeError on variable validation batch size

1

### Description Changed dimension in concat_sequences to cat along dim = 0. This fixes bug mentioned in linked issue. ### Checklist - [ ] [Linked issues (if existing)](https://github.com/jdb78/pytorch-forecasting/issues/1509) - [...

Luke-Chesley

Temporal Fusion Transformer gives constant prediction

40

- PyTorch-Forecasting version: 1.0.0 - PyTorch version: 2.0.1 - Python version: 3.10 - Operating System: Windows ### Expected behavior My expectation is to get a prediction that is not completely...

manitadayon

[ENH] Support for state of the art TSMixer model

2

I was reading [this article](https://blog.research.google/2023/09/tsmixer-all-mlp-architecture-for-time.html?m=1) on the TSMixer model, which out performs TFT. Are there any plans in the near future to incorporate TSMixer in PyTorch Forecasting? Do you know...

ruuttt

[ENH] Support tuning any model and extend LSTMModel in docs to support multi-target datasets

4

### Description This PR fixes my own issues: * https://github.com/jdb78/pytorch-forecasting/issues/1440 * https://github.com/jdb78/pytorch-forecasting/issues/1448 It allows to tune "any" model, generalizing the `optimize_hypetparameters` function from `TemporalFusionTransformer` model, and it also add a...

svnv-svsv-jm

RuntimeError: index_add(): Expected non int64 dtype for source.

6

- PyTorch-Forecasting version: 1.0.0 - PyTorch version: 2.0.1 - Python version: 3.10.10 - Operating System: macOS-13.3.1 M1 Pro - Torch.backends.mps.is_available(): True - ### Expected behavior I am trying to execute...

erl61