You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
In brief: a middleware is installed into a system, allowing programs or a user to construct neural transformations using a special API or GUI using the API and utilising unrelated models satisfying some requirements and packaged satisfying some requirements.
Is your feature request related to a problem? Please describe.
There exist 3 models for different pairs of "modalities" (see the link): A -> B and C -> D, but there exists no models for A -> D and C -> B. The middleware allows a user or a program to pretty easily get the models A -> D and C->B with possible penalty on model performance and resource consumption, but without bloat and without major cost related to full retraining from scratch.
Describe the solution you'd like
The idea of the solution is described in some detail in open-source-ideas/ideas#152 . neuropod seems to provide some elements that may be helpful for that.
Describe alternatives you've considered
Haven't considered alternatives, the alternative is to not to solve the problem at all. I assummed that the problem worth solving, but I may be wrong.
The text was updated successfully, but these errors were encountered:
Feature
In brief: a middleware is installed into a system, allowing programs or a user to construct neural transformations using a special API or GUI using the API and utilising unrelated models satisfying some requirements and packaged satisfying some requirements.
Is your feature request related to a problem? Please describe.
There exist 3 models for different pairs of "modalities" (see the link): A -> B and C -> D, but there exists no models for A -> D and C -> B. The middleware allows a user or a program to pretty easily get the models A -> D and C->B with possible penalty on model performance and resource consumption, but without bloat and without major cost related to full retraining from scratch.
Describe the solution you'd like
The idea of the solution is described in some detail in open-source-ideas/ideas#152 . neuropod seems to provide some elements that may be helpful for that.
Describe alternatives you've considered
Haven't considered alternatives, the alternative is to not to solve the problem at all. I assummed that the problem worth solving, but I may be wrong.
The text was updated successfully, but these errors were encountered: