search

SalesforceAIResearch / uni2ts

[ICML2024] Unified Training of Universal Time Series Forecasting Transformers
Apache License 2.0
611 stars 49 forks source link

Getting Started Example Bugs #34

Closed Sunsoft-io closed 1 month ago

Sunsoft-io commented 1 month ago

Hello,

I've run the Getting Started Example.

There was a torch bug, which I could resolve by installing torch 2.2.2 (instead of 2.3.0) with "pip install torch==2.2.2".

But now there is the following message: Exception ignored in: <generator object PyTorchPredictor.predict at 0x0000022E59436350> Traceback (most recent call last): File "C:\Users\willi\AppData\Roaming\Python\Python311\site-packages\gluonts\torch\model\predictor.py", line 89, in predict File "C:\Users\willi\AppData\Roaming\Python\Python311\site-packages\torch\autograd\grad_mode.py", line 84, in exitTypeError: 'NoneType' object is not callable

Thanks...Willi

gorold commented 1 month ago

Have you pulled the latest version of the repo and getting started example? Also, please provide a minimal reproducible example for this error.

Sunsoft-io commented 1 month ago

Yes I've pulled the latest version of the repo. It installed torch 2.3.0 which also didn't work.

Next I will test Lag-Lama.

Sunsoft-io commented 1 month ago

I've used the Getting Started Example of huggingface.co, which is a little bit different.

Sunsoft-io commented 1 month ago

It is different in 2 lines. When adding these 2 lines, the matplotlib graph is drawn. It seems to work!!!

But there is still the message: "TypeError: 'NoneType' object is not callable".