-
Notifications
You must be signed in to change notification settings - Fork 862
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
[timeseries] Speed up prediction for GluonTS models #2593
Conversation
Job PR-2593-4386fe1 is done. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Looks great, and very impressive speedups!! Some minor comments.
timeseries/src/autogluon/timeseries/models/gluonts/abstract_gluonts.py
Outdated
Show resolved
Hide resolved
timeseries/src/autogluon/timeseries/models/gluonts/abstract_gluonts.py
Outdated
Show resolved
Hide resolved
timeseries/src/autogluon/timeseries/models/gluonts/torch/models.py
Outdated
Show resolved
Hide resolved
4386fe1
to
ba1b94c
Compare
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM! Thanks!
Job PR-2593-6bb80b7 is done. |
Description of changes:
pd.Series
instead ofTimeSeriesDataFrame
insideSimpleGluonTSDataset
to avoid copying the static features.get_forecast_horizon_index_ts_dataframe
with a more efficientgroupby
operationAbstractGluonTSModel. predict
.Benchmarking results on the M5 dataset
(only training for 1 epoch / 1 batch, so the training time/performance are meaningless here)
With this PR
5K items / 7M rows
30K items / 47M rows
Current
master
branch5K items / 7M rows
30K items / 47M rows
Code to reproduce the results
By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.