Differentiable Architecture Search (DART) is a method for efficient architecture search. The search space is made continuous so that the architecture can be optimized with respect to its validation set performance through gradient descent.
Source: DARTS: Differentiable Architecture SearchPaper | Code | Results | Date | Stars |
---|
Task | Papers | Share |
---|---|---|
Image Classification | 15 | 12.82% |
Reinforcement Learning (RL) | 6 | 5.13% |
Semantic Segmentation | 5 | 4.27% |
Bilevel Optimization | 5 | 4.27% |
Language Modelling | 5 | 4.27% |
General Classification | 5 | 4.27% |
Network Pruning | 4 | 3.42% |
Classification | 4 | 3.42% |
BIG-bench Machine Learning | 4 | 3.42% |
Component | Type |
|
---|---|---|
🤖 No Components Found | You can add them if they exist; e.g. Mask R-CNN uses RoIAlign |