Giters

scorelab / tensormap

TensorMap will be a web application that will allow the users to create machine learning algorithms visually. TensorMap will support reverse engineering of the visual layout to a Tensorflow implementation in preferred languages. The goal of the project is to let the beginners play with machine learning algorithms in Tensorflow without less background knowledge about the library.

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Execution Configuration Dialog

vasudev13 opened this issue · comments

commented

To make things easier and avoid confusion, we can have selects for configuration settings. Moreover, while defining the Optimizer, there should be an number input for Learning Rate.

Chip Multiselect was used for metrics as having multiple metrics is common while evaluating a network(though same can be achieved using checkboxes) . Plus, if an additional functionality is to have Custom Metrics (very common in ML Competitions), allowing user to define an custom python function in a textarea.

Peek 2020-03-24 14-17

PR will be established, if this fits well within the application's design and intuition.

commented

@vasudev13 Do check the backend if it contain such options :)

commented

As in keras's backend? or application's backend. I believe earlier options were hard-coded in config object, I have only changed the UI elements. I do agree a list can be populated from a API, but I don't think any such implementation is present as of now. I could be wrong.