automated-machine-learning

Feature Description

We want to enable the users to specify the value ranges for any argument in the blocks.
The following code example shows a typical use case.
The users can specify the number of units in a DenseBlock to be either 10 or 20.

Code Example

import au

It seems there is no validation on fit_ensemble when ensemble size is 0, causing an issue to appear as seen in #1327

• transform_primitive.py is becoming very large. I suggest splitting out into separate files.
• We could split this file up by groups (such as LatLong transform primitives in 1 file).

Related: awslabs/autogluon#1479

Add a scikit-learn compatible API wrapper of TabularPredictor:

• TabularClassifier
• TabularRegressor

Required functionality (may need more than listed):

• init API
• fit API
• predict API
• works in sklearn pipelines

Problem
Some of our transformers & estimators are not thoroughly tested or not tested at all.

Solution
Use OpTransformerSpec and OpEstimatorSpec base test specs to provide tests for all existing transformers & estimators.

This issue has been coming up when I use,
automl.predict_proba(input)

I am using the requirements.txt in venv. Shouldn't input have feature names?
This message did not used to come up and I don't know why.

In principle it seems getting the parameters from FLAML to C# LightGBM seems to work, but I dont have any metrics yet. The names of parameters are slightly different but documentation is adequate to match them. Microsoft.ML seems to have version 2.3.1 of LightGBM.

Another approach that might be useful, especially for anyone working with .NET, would be having some samples about conversion to ONN

In tuner_interface.py function _is_greater_better does not guarantee correct output.

Example 1:
The area under the precision-recall curve for unbalanced data might be less than 0,5 and the function will return False, while metrics needs to be maximized

Example 2:
Custom metrics that needs to be minimized is not obliged to converge to zero

It would help to have download option to get a list of used packages:

For example:
Download list of ML Packages used in Model Training, with all corresponding citations in CSV Format.