Wu Yi
Wu Yi
 
In the refactored code, we use Couler to generate an Argo YAML file to submit to the Kubernetes cluster to run. A generated Couler program should look like: ```python def...
The current `sqlflow release repo ...` command does not output kaniko build log, the error message can only be viewed on the server-side. We should pipe the kaniko log to...
Start sqlflowserver in the Docker container by `./sqlflowserver &`, then run SQL statement ``` create table mytable as select ...; ``` if `mytable` is already exists, the below error is...
Generally, as a compiler, SQLFlow is performing the below steps to compile the SQL program to a workflow, the workflow is usually a `.yaml` file: ``` parse -> AST ->...
The `TO EVALUATE` statement should output metrics for each sample for later data visualization.
Currently, the SQLFlow Python runtime library rederives features (decide how to convert table fields into model input tensors) in each iteration of the train loop. https://github.com/sql-machine-learning/sqlflow/blob/957ac02be6737a1a25058ba6454f3c565fbb59bb/python/sqlflow_submitter/tensorflow/input_fn.py#L22-L32 This is unnecessary; instead,...
- [x] save the model on Maxcompute using `sqlfs` package so that the SQLflow cli can download the model and do release. - [x] the deployed server should support `docker...
We should output parse result without comment lines to make later processing easier. Currently, below SQL script: ```sql -- some comment set key=value; -- some comment SELECT * FROM table;...
Please edit this issue to add release notes below: - Support `TO EVALUATE` clause to evaluate a model. - SQLFlow model zoo, support publicly share model definitions and models.