PUBDEV-8047: AutoML Save/Load implementation #6437
Open
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
tomasfryda
commented
Nov 25, 2022
tomasfryda
commented
Nov 25, 2022
Comment on lines
959
to
984
| protected static AutoML readAutoML(AutoBuffer ab, Futures fs) { | ||
| try (PersistenceContext pc = PersistenceContext.begin()) { | ||
| AutoML aml = new AutoML(ab.get(), null, ab.get(), false); | ||
| aml._leaderboard = (Leaderboard) ab.getKey(fs); | ||
| aml._eventLog = (EventLog) ab.getKey(fs); | ||
| // aml._trainingFrame = (Frame) ab.getKey(fs); | ||
| fs.blockForPending(); | ||
| for (Key mk : aml.leaderboard().getModelKeys()) { | ||
| Model m = (Model) PersistenceContext.getKey(ab, fs, mk); | ||
| if (aml._buildSpec.build_control.keep_cross_validation_predictions) | ||
| for (Key k : m._output._cross_validation_predictions) | ||
| PersistenceContext.loadKey(ab, fs, k); | ||
| if (aml._buildSpec.build_control.keep_cross_validation_models) | ||
| for (Key k : m._output._cross_validation_models) | ||
| PersistenceContext.loadKey(ab, fs, k); | ||
| if (aml._buildSpec.build_control.keep_cross_validation_fold_assignment) | ||
| PersistenceContext.loadKey(ab,fs ,m._output._cross_validation_fold_assignment_frame_id); | ||
| // if (m instanceof StackedEnsembleModel) | ||
| // PersistenceContext.loadKey(ab, fs, ((StackedEnsembleModel)m)._output._metalearner._parms._train); | ||
| } | ||
| DKV.put(aml); | ||
| return aml; | ||
| } catch (Exception e) { | ||
| throw new RuntimeException(e); | ||
| } | ||
| } |
Contributor
Author
There was a problem hiding this comment.
The objects have to be deserialized exactly in the same order as they were serialized => we need to keep track of what was already loaded and in that case skip the loading of that object.
tomasfryda
force-pushed
the
tomf_PUBDEV-8047_save_load_automl
branch
2 times, most recently
from
d4fb774 to
5ba924b
Compare
tomasfryda
force-pushed
the
tomf_PUBDEV-8047_save_load_automl
branch
from
5ba924b to
74490f2
Compare
ledell
approved these changes
Jan 10, 2023
Contributor
ledell
left a comment
ledell
left a comment
There was a problem hiding this comment.
This looks great, thank you!
|
. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
https://h2oai.atlassian.net/browse/PUBDEV-8047
Known issues:
automl_training_frameget_leaderboard("ALL")on loaded object producesNAs forpredict_time_per_row_mson loaded modelspredict_time_per_row_msis successfully saved and loaded.make_leaderboard("ALL")on loaded object without providedleaderboard_framefails onpredict_time_per_row_ms