search

time-series-foundation-models / lag-llama

Lag-Llama: Towards Foundation Models for Probabilistic Time Series Forecasting
Apache License 2.0
1.08k stars 121 forks source link

item_id error #54

Open drorhunvural opened 2 months ago

drorhunvural commented 2 months ago

Hi,

I am running Colab Demo 1 using my own dataset, which is structured on an hourly basis, similar to the one shown below.

image

I'm trying to predict "count." Since my dataset doesn't contain multiple series, I believe I don't need an item_id, right?

If that's the case, how can I omit the item_id in the code below, as I'm encountering an error related to it?

dataset = PandasDataset.from_long_dataframe(df, target="count")

Error

image

ashok-arjun commented 2 months ago

Hi,

This is an issue with the GluonTS class itself; it needs an item_id nevertheless, even if it's just one series. Can you try adding an item_id column in your dataset, with a random value for it?