Hydra is a multi-headed neural network for model distillation with a shared body network. The shared body network learns a joint feature representation that enables each head to capture the predictive behavior of each ensemble member. Existing distillation methods often train a distillation network to imitate the prediction of a larger network. Hydra instead learns to distill the individual predictions of each ensemble member into separate light-weight head models while amortizing the computation through a shared heavy-weight body network. This retains the diversity of ensemble member predictions which is otherwise lost in knowledge distillation.
Source: Hydra: Preserving Ensemble Diversity for Model DistillationPaper | Code | Results | Date | Stars |
---|
Task | Papers | Share |
---|---|---|
Language Modelling | 2 | 18.18% |
Earth Observation | 1 | 9.09% |
Virtual Try-on | 1 | 9.09% |
Management | 1 | 9.09% |
Benchmarking | 1 | 9.09% |
Recommendation Systems | 1 | 9.09% |
Experimental Design | 1 | 9.09% |
Decoder | 1 | 9.09% |
Domain Adaptation | 1 | 9.09% |
Component | Type |
|
---|---|---|
🤖 No Components Found | You can add them if they exist; e.g. Mask R-CNN uses RoIAlign |