Innixma
Follow
Pinned
659 contributions in the last year
Less
More
Contribution activity
March 2021
Created 8 commits in 2 repositories
Created a pull request in awslabs/autogluon that received 10 comments
Tabular: Added LOO-based bagging alternative to KNN
Issue #, if available: Description of changes: Added LOO-based bagging alternative to KNN (enabled by default). Now, during bagging KNN will only t…
+383
−50
•
10
comments
Opened 6 other pull requests in 1 repository
Reviewed 15 pull requests in 2 repositories
awslabs/autogluon 14 pull requests
- Constrained Bayesian optimization
- Fix autogluon.mxnet crash on submodule install
- Tabular: Added LOO-based bagging alternative to KNN
- [Tabular] initial Quantile Regression
- [DO-NOT-MERGE-YET][Dependencies] CUDA compatibility issues in linux
- [Vision] Tutorials for loading data
- [Schedulers] Schedulers Refactoring
- Fixed exception when providing a dataset to ENAS_Scheduler train_set …
- add sample_weight to score_debug EnsembleSelection
- Tabular: Added initial RAPIDS KNN model
- Tabular: Added initial RFRapidsModel
- [CI] Fix torchvision dataset url
- Tabular: Add option to calculate feature group importance
- Update hpo tutorial to remove random search + hyperband opt because they are not stable currently
aws/amazon-sagemaker-examples 1 pull request
Created an issue in awslabs/autogluon that received 4 comments
[Vision] Exception when loading GPU trained model on CPU machine
When an ImagePredictor object is trained on GPU, then loaded on a CPU-only machine for inference, this error occurs:
MXNetError: GPU is not enabled
I…
4
comments
Opened 14 other issues in 2 repositories
awslabs/autogluon
12
open
1
closed
- Child model leaderboard visibility + variance calculation
- BLAS : Program is Terminated. Because you tried to allocate too many memory regions.
- Improve logging clarity of TextPredictor exceptions when called via TabularPredictor
- Add model uncertainty as feature to stack when performing repeated bagging.
- Clarify pip install autogluon.tabular[all] instructions
- Error when calling leaderboard after using high_quality preset with TextPredictor model.
- Fix fit_weighted_ensemble on high_quality preset
- Document submodule installation
- Tabular: Add option for user to specify # of folds to use during inference
- Fix TextPredictorModel during refit_full
- Add log statement that eval_metric scores are being flipped to ensure higher is better.
- RF/XT lower priority for regression problems.
- Add feature importance groupings to Tabular
openml/automlbenchmark
1
open
Answered 1 discussion in 1 repository
awslabs/autogluon
awslabs/autogluon