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] Update documentation #3297
Conversation
Check out this pull request on See visual diffs & provide feedback on Jupyter Notebooks. Powered by ReviewNB |
Job PR-3297-1c351d7 is done. |
Please refer to | ||
the GluonTS [documentation](https://ts.gluon.ai/stable/api/gluonts/gluonts.html) and | ||
[github](https://github.com/awslabs/gluon-ts) for further information. | ||
Currently, AutoGluon does not support features such as hierarchical forecasting and forecast explainability, but we will consider adding them in the future. |
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.
Is there an issue with tracking feature requests related to hierarchical forecasting and explainability from users? If yes, we can remove the "but" statement.
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.
Just to clarify, you think we should open an issue for each of these features on GitHub & link them here so that users can "+1" if they care about these?
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.
I support adding feature issues to our backlog and linking them to release milestones to ensure we stay on track. Additionally, if users have already requested certain features, we can include that information in the issue description.
|
||
|
||
## Can I use GPUs for model training? | ||
|
||
Yes! Most of the deep learning models used by `autogluon.timeseries` support GPU training. | ||
PyTorch models will have GPU enabled by default. If you also want to use MXNet models, make sure you have installed CUDA and the GPU version of MXNet. | ||
Yes! All deep learning models used by `autogluon.timeseries` support GPU training. |
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.
Can we add some instructions to help users understand if GPU is used/supported for TS training? For example, specific log entry emitted
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.
Currently this would only be visible for verbosity=4
(in the pytorch-lightning device information), I don't think that's recommended. We could add the following log entry inside TimeSeriesPredictor.fit
if torch.cuda.is_available():
logger.info("GPU detected. Models that support GPU training will be trained on GPU")
Are you in favor of doing this?
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.
Do we have log entry from lightning such as:
Using 16bit None Automatic Mixed Precision (AMP)
GPU available: True (cuda), used: True
TPU available: False, using: 0 TPU cores
IPU available: False, using: 0 IPUs
HPU available: False, using: 0 HPUs
LOCAL_RANK: 0 - CUDA_VISIBLE_DEVICES: [0]
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.
These only appear at verbosity=4
together with the rest of the GluonTS output (which can be very crowded). I wouldn't recommend that users set verbosity=4
unless debugging.
Job PR-3297-ee4b05a is done. |
Job PR-3297-94f6019 is done. |
* 'master' of https://github.com/awslabs/autogluon: (24 commits) [WIP] 0.8.0 release notes (autogluon#3303) Add model keys doc (autogluon#3321) Fix NaN warning in np.array(X) (autogluon#3315) [Draft] Upgrade networkx to 3.x (autogluon#3317) Add calibrate_decision_threshold tutorial (autogluon#3316) [Doc] AutoMM FAQ Updates (autogluon#3314) Update to v0.8.0 (autogluon#3313) Add Experimental Zeroshot HPO (autogluon#3312) Update GPU installation guide to use CUDA 11.7 (autogluon#3306) [Tutorial]Update tutorials for object detection (autogluon#3305) [timeseries] Update documentation (autogluon#3297) Update mac cpu install instructions (autogluon#3280) Add docstring for hyperparameter_tune_kwargs (autogluon#3307) [Doc] Add Search Space Page (autogluon#3311) Fewshot learning predict proba (autogluon#3267) Fix log to file Windows tests (autogluon#3302) Add missing doc pages (autogluon#3304) Add calibrate_decision_threshold (autogluon#3298) continuous training tutorial update (autogluon#3300) Add TabPFN (autogluon#3270) ...
Description of changes:
To do:
By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.