sqlflow icon indicating copy to clipboard operation
sqlflow copied to clipboard

Brings SQL and AI together.

Results 107 sqlflow issues
Sort by recently updated
recently updated
newest added

After our code refactory, only workflow mode is officially supported. This makes the developing and debugging more complicated. The complication comes from ways: 1. Many docker images (dev, ci, server,...

doc
Refactory

**Is your feature request related to a problem? Please describe.** A clear description of the problem **Describe the solution you'd like** A clear description of what you want to happen...

**Is your feature request related to a problem? Please describe.** As described in [language guide](https://github.com/sql-machine-learning/sqlflow/blob/develop/doc/language_guide.md#to-run-syntax), we have some pre-made runnable in sqlflow/runnable, but there is no doc now, please add...

doc

Users from community asked about how to build a model in recommendation scenarios. It would be better that we can provide a tutorial of CTR model in the tutorial.

**Description** A clear description of the bug **Reproduction Steps** Detailed steps to reproduce **Expected Behavior** What you expected to happen. **Screenshots** **Environment (Please complete the following information):** - OS: -...

**Is your feature request related to a problem? Please describe.** Our language guide mentions the models parameter and hyperparameter, like [here](https://github.com/sql-machine-learning/sqlflow/blob/develop/doc/language_guide.md#train-clause), but users has nowhere to refer to them. **Describe...

doc

**Is your feature request related to a problem? Please describe.** Add a doc to describe how to use our explainer. Then link it to our [language guide](https://github.com/sql-machine-learning/sqlflow/blob/develop/doc/language_guide.md#explain-and-using-clause) **Describe the solution...

doc

**Is your feature request related to a problem? Please describe.** Currently there is no doc for models we used in the example. We should find a place to include generic...

feature

At the present, TO EXPLAIN for XGBoost model only support single machine execution. It will cost a long time for a large dataset. Distributed execution is necessary for performance speed...