Timeline for Is there a method to apply trained weights to a model with the same input shape and model architecture but different output shape?
Current License: CC BY-SA 4.0
3 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Jul 9, 2020 at 15:50 | comment | added | Nikos M. | Why this would work? What is the underlying principle that it is based on? | |
| Jul 9, 2020 at 15:47 | comment | added | Nikos M. | And how are these 15 neurons of the new layer supposed to be adjusted (trained)? And if trained over the data what is the advantage over training a new neural net from scratch that classifies the new problem directly? | |
| Jul 9, 2020 at 11:28 | history | answered | 10xAI | CC BY-SA 4.0 |