Picture for Vinod Ganesan

Vinod Ganesan

SuperShaper: Task-Agnostic Super Pre-training of BERT Models with Variable Hidden Dimensions

Add code
Oct 10, 2021
Figure 1 for SuperShaper: Task-Agnostic Super Pre-training of BERT Models with Variable Hidden Dimensions
Figure 2 for SuperShaper: Task-Agnostic Super Pre-training of BERT Models with Variable Hidden Dimensions
Figure 3 for SuperShaper: Task-Agnostic Super Pre-training of BERT Models with Variable Hidden Dimensions
Figure 4 for SuperShaper: Task-Agnostic Super Pre-training of BERT Models with Variable Hidden Dimensions
Viaarxiv icon

Design and Scaffolded Training of an Efficient DNN Operator for Computer Vision on the Edge

Add code
Aug 25, 2021
Figure 1 for Design and Scaffolded Training of an Efficient DNN Operator for Computer Vision on the Edge
Figure 2 for Design and Scaffolded Training of an Efficient DNN Operator for Computer Vision on the Edge
Figure 3 for Design and Scaffolded Training of an Efficient DNN Operator for Computer Vision on the Edge
Figure 4 for Design and Scaffolded Training of an Efficient DNN Operator for Computer Vision on the Edge
Viaarxiv icon

FuSeConv: Fully Separable Convolutions for Fast Inference on Systolic Arrays

Add code
May 27, 2021
Viaarxiv icon