Picture for Elnur Gasanov

Elnur Gasanov

Error Feedback Reloaded: From Quadratic to Arithmetic Mean of Smoothness Constants

Feb 16, 2024
Figure 1 for Error Feedback Reloaded: From Quadratic to Arithmetic Mean of Smoothness Constants
Figure 2 for Error Feedback Reloaded: From Quadratic to Arithmetic Mean of Smoothness Constants
Figure 3 for Error Feedback Reloaded: From Quadratic to Arithmetic Mean of Smoothness Constants
Figure 4 for Error Feedback Reloaded: From Quadratic to Arithmetic Mean of Smoothness Constants
Viaarxiv icon

Understanding Progressive Training Through the Framework of Randomized Coordinate Descent

Add code
Jun 06, 2023
Figure 1 for Understanding Progressive Training Through the Framework of Randomized Coordinate Descent
Figure 2 for Understanding Progressive Training Through the Framework of Randomized Coordinate Descent
Figure 3 for Understanding Progressive Training Through the Framework of Randomized Coordinate Descent
Figure 4 for Understanding Progressive Training Through the Framework of Randomized Coordinate Descent
Viaarxiv icon

Error Feedback Shines when Features are Rare

Add code
May 24, 2023
Figure 1 for Error Feedback Shines when Features are Rare
Figure 2 for Error Feedback Shines when Features are Rare
Figure 3 for Error Feedback Shines when Features are Rare
Figure 4 for Error Feedback Shines when Features are Rare
Viaarxiv icon

Adaptive Compression for Communication-Efficient Distributed Training

Oct 31, 2022
Figure 1 for Adaptive Compression for Communication-Efficient Distributed Training
Figure 2 for Adaptive Compression for Communication-Efficient Distributed Training
Figure 3 for Adaptive Compression for Communication-Efficient Distributed Training
Figure 4 for Adaptive Compression for Communication-Efficient Distributed Training
Viaarxiv icon

3PC: Three Point Compressors for Communication-Efficient Distributed Training and a Better Theory for Lazy Aggregation

Feb 02, 2022
Figure 1 for 3PC: Three Point Compressors for Communication-Efficient Distributed Training and a Better Theory for Lazy Aggregation
Figure 2 for 3PC: Three Point Compressors for Communication-Efficient Distributed Training and a Better Theory for Lazy Aggregation
Figure 3 for 3PC: Three Point Compressors for Communication-Efficient Distributed Training and a Better Theory for Lazy Aggregation
Figure 4 for 3PC: Three Point Compressors for Communication-Efficient Distributed Training and a Better Theory for Lazy Aggregation
Viaarxiv icon

FLIX: A Simple and Communication-Efficient Alternative to Local Methods in Federated Learning

Add code
Nov 22, 2021
Figure 1 for FLIX: A Simple and Communication-Efficient Alternative to Local Methods in Federated Learning
Figure 2 for FLIX: A Simple and Communication-Efficient Alternative to Local Methods in Federated Learning
Figure 3 for FLIX: A Simple and Communication-Efficient Alternative to Local Methods in Federated Learning
Figure 4 for FLIX: A Simple and Communication-Efficient Alternative to Local Methods in Federated Learning
Viaarxiv icon

Lower Bounds and Optimal Algorithms for Smooth and Strongly Convex Decentralized Optimization Over Time-Varying Networks

Jun 08, 2021
Figure 1 for Lower Bounds and Optimal Algorithms for Smooth and Strongly Convex Decentralized Optimization Over Time-Varying Networks
Figure 2 for Lower Bounds and Optimal Algorithms for Smooth and Strongly Convex Decentralized Optimization Over Time-Varying Networks
Figure 3 for Lower Bounds and Optimal Algorithms for Smooth and Strongly Convex Decentralized Optimization Over Time-Varying Networks
Figure 4 for Lower Bounds and Optimal Algorithms for Smooth and Strongly Convex Decentralized Optimization Over Time-Varying Networks
Viaarxiv icon

From Local SGD to Local Fixed Point Methods for Federated Learning

Add code
Apr 03, 2020
Figure 1 for From Local SGD to Local Fixed Point Methods for Federated Learning
Figure 2 for From Local SGD to Local Fixed Point Methods for Federated Learning
Figure 3 for From Local SGD to Local Fixed Point Methods for Federated Learning
Figure 4 for From Local SGD to Local Fixed Point Methods for Federated Learning
Viaarxiv icon