Ensembling

Multi-Input Multi-Output training of neural networks
Multi-Input Multi-Output training leads to the emergence of smaller subnetworks within the base network. This has a number of benefits such a better utilization of network parameters, better training and strong regularization of learned features.
Reconciling feature sharing and multiple predictions with MIMO Vision Transformers
Multi-input multi-output training improves network performance by optimizing multiple subnetworks simultaneously. In this paper, we …
Remy Sun, Clément Masson, Nicolas Thome, Matthieu Cord
PDF Cite Project Project
Reconciling feature sharing and multiple predictions with MIMO Vision Transformers
Adapting Multi-input Multi-output Schemes to Vision Transformers
Multi-input multi-output models have proven capable of providing ensembling for free in convolutional neural networks by training …
Remy Sun, Clément Masson, Nicolas Thome, Matthieu Cord
PDF Cite Project Project
Adapting Multi-input Multi-output Schemes to Vision Transformers
Towards Efficient Feature Sharing in MIMO Architectures
Multi-input multi-output architectures propose to train multiple subnetworks within one base network and then average the subnetwork …
Remy Sun, Alexandre Rame, Clément Masson, Nicolas Thome, Matthieu Cord
PDF Cite Project
Towards Efficient Feature Sharing in MIMO Architectures
MixMo Mixing Multiple Inputs for Multiple Outputs via Deep Subnetworks
Recent strategies achieved ensembling “for free” by fitting concurrently diverse subnetworks inside a single base network. …
Alexandre Rame, Remy Sun, Matthieu Cord
PDF Cite Code Project Project
MixMo Mixing Multiple Inputs for Multiple Outputs via Deep Subnetworks