Document current limitation in number of features #3886
Merged
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
|
@trivialfis FYI |
|
@hcho3 What's preventing using int64_t currently? |
|
@trivialfis XGBoost saves the trained model as binary dump. So you can't easily add or change fields without breaking backward compatibility. As an aside, someone told me today that the use of binary dump is causing endian-compatibility issue as well (model created by little-endian machine doesn't load on big-endian machine). We may want to consider a next generation model format. |
|
@trivialfis Since dense structures are currently used in many places to index features, using such large indices would be consuming many gigs of memory. |
This was referenced Nov 12, 2018
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
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Outdated suggestions cannot be applied.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.
See discussion at #3884. The
num_featurevariable uses 32-bit integer, which causes XGBoost to fail for datasets with more than 2147483647 features. For now, print a better error message.