In a completely different field with a completely different angle, I've been working on an almost identical application. The commonality is that both are 'just' graphical pipeline editors where I have been focussing on neuroimaging and MRI data analysis, you have been focussing on machine/deep learning.
My focus on reproducible analysis is similar and the application is called: Graphical Interface for Reproducible Analysis oF workFlow Experiments, https://giraffe.tools. I looked through your code and found that the core is similar (Django + React) but the application is designed completely differently. I loved your layout design, so I recently copied some of that and thus may look familiar. Here's an example: https://giraffe.tools/porcupine/TimVanMourik/SomeGiraffeExample.
Some of the biggest differences with my platform
Social OAuth implemented with GitHub
Read from and write to GitHub implemented
D3 based UI (svg) instead of html
s
React-Redux state management
Radium based styling instead of CSS
Code representation of the workflow (graph net) is shown
The libraries of nodes is dynamically loaded instead of hard-coded
Your graphic design was a lot better. I borrowed some of the styling for my own app, with gratitude
I'd love to see if there's some way in which we could help each other out. I've been thinking for a while to connect my rather generic workflow package to machine learning tools. For instance I would love to create TensorFlow/Keras/Pytorch code with GiraffeTools out of the box!
Collaboration possibilities
In a completely different field with a completely different angle, I've been working on an almost identical application. The commonality is that both are 'just' graphical pipeline editors where I have been focussing on neuroimaging and MRI data analysis, you have been focussing on machine/deep learning.
My focus on reproducible analysis is similar and the application is called: Graphical Interface for Reproducible Analysis oF workFlow Experiments, https://giraffe.tools. I looked through your code and found that the core is similar (Django + React) but the application is designed completely differently. I loved your layout design, so I recently copied some of that and thus may look familiar. Here's an example: https://giraffe.tools/porcupine/TimVanMourik/SomeGiraffeExample.
Some of the biggest differences with my platform
I'd love to see if there's some way in which we could help each other out. I've been thinking for a while to connect my rather generic workflow package to machine learning tools. For instance I would love to create TensorFlow/Keras/Pytorch code with GiraffeTools out of the box!