Picture for Marius Preda

Marius Preda

TSP, IP Paris, SAMOVAR

Induced Feature Selection by Structured Pruning

Add code
Mar 20, 2023
Viaarxiv icon

Improve Convolutional Neural Network Pruning by Maximizing Filter Variety

Add code
Mar 11, 2022
Figure 1 for Improve Convolutional Neural Network Pruning by Maximizing Filter Variety
Figure 2 for Improve Convolutional Neural Network Pruning by Maximizing Filter Variety
Figure 3 for Improve Convolutional Neural Network Pruning by Maximizing Filter Variety
Figure 4 for Improve Convolutional Neural Network Pruning by Maximizing Filter Variety
Viaarxiv icon

An Experimental Study of the Impact of Pre-training on the Pruning of a Convolutional Neural Network

Add code
Dec 15, 2021
Figure 1 for An Experimental Study of the Impact of Pre-training on the Pruning of a Convolutional Neural Network
Figure 2 for An Experimental Study of the Impact of Pre-training on the Pruning of a Convolutional Neural Network
Figure 3 for An Experimental Study of the Impact of Pre-training on the Pruning of a Convolutional Neural Network
Figure 4 for An Experimental Study of the Impact of Pre-training on the Pruning of a Convolutional Neural Network
Viaarxiv icon

One-Cycle Pruning: Pruning ConvNets Under a Tight Training Budget

Add code
Jul 05, 2021
Figure 1 for One-Cycle Pruning: Pruning ConvNets Under a Tight Training Budget
Figure 2 for One-Cycle Pruning: Pruning ConvNets Under a Tight Training Budget
Figure 3 for One-Cycle Pruning: Pruning ConvNets Under a Tight Training Budget
Figure 4 for One-Cycle Pruning: Pruning ConvNets Under a Tight Training Budget
Viaarxiv icon

End-to-end deep metamodeling to calibrate and optimize energy loads

Add code
Jun 19, 2020
Figure 1 for End-to-end deep metamodeling to calibrate and optimize energy loads
Figure 2 for End-to-end deep metamodeling to calibrate and optimize energy loads
Figure 3 for End-to-end deep metamodeling to calibrate and optimize energy loads
Figure 4 for End-to-end deep metamodeling to calibrate and optimize energy loads
Viaarxiv icon