Giters
mindsdb
/
lightwood
Lightwood is Legos for Machine Learning.
Geek Repo:
Geek Repo
Github PK Tool:
Github PK Tool
Stargazers:
436
Watchers:
17
Issues:
453
Forks:
92
mindsdb/lightwood Issues
Refactor setup to use Poetry instead
Closed
4 months ago
Comments count
1
Pin all requirements
Closed
4 months ago
Re-design accuracy functions
Updated
5 months ago
Comments count
1
Allow to `finetune` GPU-trained models on CPU
Updated
6 months ago
Comments count
1
BestOf ensemble: option to specify whether reported metric should be minimized or maximized
Updated
6 months ago
[docs] Optional classes are not documented
Updated
6 months ago
Migrations support for lightwood
Updated
6 months ago
Add hyperparam search to XGBoost mixer
Updated
6 months ago
Add partial fit support for XGBoost mixer
Updated
6 months ago
Add array-compatible XGBoost and RandomForest mixers
Updated
6 months ago
Comments count
2
Remove old code (rnn_helpers and related)
Closed
6 months ago
Comments count
1
Best Of ensemble: solve mixer ties with runtime as second criterion
Closed
6 months ago
Comments count
4
Improve release.yml
Closed
6 months ago
Comments count
5
Deprecate manual offset to forecasts
Closed
6 months ago
Comments count
1
error when HORIZON=1
Closed
6 months ago
Comments count
8
Benchmark datetime encoders
Closed
6 months ago
Comments count
1
Add forecasting support with XGBoost
Closed
10 months ago
Comments count
2
Fix 'ambiguous variable name 'l' issue in lightwood\encoder\text\helpers\rnn_helpers.py
Closed
a year ago
Comments count
2
[StatsForecast] ValueError: Input contains NaN, infinity or a value too large for dtype('float64')
Closed
a year ago
Comments count
2
Support custom nltk paths
Closed
a year ago
Handle timeseries problems with gaps and/or inconsistent frequency
Closed
a year ago
Comments count
2
Investigate how a user can get single predictions on timeseries models
Closed
2 years ago
Comments count
4
Add all columns from Prophet mixer in explain output
Closed
a year ago
[Bug, ML]: unhashable type: 'numpy.ndarray' on timeseries data
Closed
a year ago
Comments count
7
Handle no input features gracefully.
Closed
a year ago
Restore custom cleaner tutorial
Closed
a year ago
Comments count
1
Bring back multiple output support
Closed
a year ago
Vectorize DatetimeNormalizerEncoder
Updated
a year ago
Change categorical AE to use label encoding instead of OHE
Closed
a year ago
Comments count
1
Scope a more intuitive declarative structure
Updated
a year ago
Comments count
1
Error: AttributeError: module 'lightwood' has no attribute 'data'
Closed
a year ago
Comments count
3
Avoid lists as cells in TS transformations
Closed
a year ago
Ensure tests pass without torch multiprocessing
Closed
a year ago
Comments count
2
Check potential issue with ignored forced dtypes in `json_ai_from_problem`
Closed
a year ago
Comments count
1
Automatically adding stop_after argument
Closed
a year ago
Automatic model weight device transfer
Updated
a year ago
Support for Python >3.10
Closed
a year ago
Comments count
2
Better error messages when target column dtype is not supported
Updated
a year ago
Comments count
7
[Bug] ICP softmax type error
Closed
a year ago
Comments count
1
Add main module to the sample code in the documentation
Closed
a year ago
Comments count
4
Improve error message if no mixers could be trained
Closed
a year ago
Comments count
1
More user-friendly message when not able to download NLTK data
Closed
a year ago
Improve ProphetMixer customization options
Closed
a year ago
Mark GluonTSMixer as optional
Closed
2 years ago
Comments count
1
Expose global insights generated in the analysis phase
Closed
2 years ago
XGBoost mixer
Closed
2 years ago
Comments count
1
Improve time series frequency inference
Closed
2 years ago
Check window transformations for time series
Closed
2 years ago
Comments count
2
Lightwood pip packages creates tests module
Closed
2 years ago
Ignore inferred dtype if encoder is manually provided.
Updated
2 years ago
Previous
Next