epfml / powersgdLinks
Practical low-rank gradient compression for distributed optimization: https://arxiv.org/abs/1905.13727
☆149Updated last year
Alternatives and similar repositories for powersgd
Users that are interested in powersgd are comparing it to the libraries listed below
Sorting:
- GRACE - GRAdient ComprEssion for distributed deep learning☆139Updated last year
- ☆46Updated 5 years ago
- Implementation of (overlap) local SGD in Pytorch☆34Updated 5 years ago
- Stochastic Gradient Push for Distributed Deep Learning☆170Updated 2 years ago
- Sketched SGD☆28Updated 5 years ago
- ☆94Updated 3 years ago
- PipeTransformer: Automated Elastic Pipelining for Distributed Training of Large-scale Models. ICML 2021☆55Updated 4 years ago
- Atomo: Communication-efficient Learning via Atomic Sparsification☆28Updated 7 years ago
- [ICLR 2018] Deep Gradient Compression: Reducing the Communication Bandwidth for Distributed Training☆226Updated last year
- Sparsified SGD with Memory: https://arxiv.org/abs/1809.07599☆58Updated 7 years ago
- ☆77Updated 6 years ago
- Efficient reference implementations of the static & dynamic M-FAC algorithms (for pruning and optimization)☆17Updated 3 years ago
- ☆222Updated 2 years ago
- Code for the signSGD paper☆92Updated 5 years ago
- Research and development for optimizing transformers☆131Updated 4 years ago
- Accuracy 77%. Large batch deep learning optimizer LARS for ImageNet with PyTorch and ResNet, using Horovod for distribution. Optional acc…☆38Updated 4 years ago
- Simple Distributed Deep Learning on TensorFlow☆134Updated 7 months ago
- Implementation of Parameter Server using PyTorch communication lib☆43Updated 6 years ago
- QSGD-TF☆21Updated 6 years ago
- Decentralized SGD and Consensus with Communication Compression: https://arxiv.org/abs/1907.09356☆75Updated 5 years ago
- ☆22Updated 5 years ago
- ☆43Updated 3 years ago
- Training neural networks in TensorFlow 2.0 with 5x less memory☆137Updated 3 years ago
- Partial implementation of paper "DEEP GRADIENT COMPRESSION: REDUCING THE COMMUNICATION BANDWIDTH FOR DISTRIBUTED TRAINING"☆31Updated 5 years ago
- [IJCAI2023] An automated parallel training system that combines the advantages from both data and model parallelism. If you have any inte…☆52Updated 2 years ago
- Block-sparse primitives for PyTorch☆160Updated 4 years ago
- gTop-k S-SGD: A Communication-Efficient Distributed Synchronous SGD Algorithm for Deep Learning☆37Updated 6 years ago
- FTPipe and related pipeline model parallelism research.☆44Updated 2 years ago
- ☆115Updated last year
- Code associated with the paper **Fine-tuning Language Models over Slow Networks using Activation Compression with Guarantees**.☆28Updated 2 years ago