1 repo
Awesome GitHub RepositoriesExpert Parallelism Configurations
Tools for distributing model parameters across process groups to manage expert-based parallelism in sparse neural networks.
Distinguishing note: Focuses on the configuration and distribution of model layers rather than the underlying execution kernels.
Explore 1 awesome GitHub repository matching artificial intelligence & ml · Expert Parallelism Configurations. Refine with filters or upvote what's useful.
Awesome Expert Parallelism Configurations GitHub Repositories
Find the best repos with AI.We'll search the best matching repositories with AI.
- deepspeedai/DeepSpeed
deepspeedai/DeepSpeed
41,638View on GitHubDeepSpeed is a high-performance library designed to scale deep learning model training and inference across massive clusters of GPUs and compute nodes. It provides a comprehensive suite of tools for distributed training, enabling the execution of models that exceed the memory capacity of single devices through advanced parameter partitioning, pipeline-based model parallelism, and memory-efficient state offloading. The framework distinguishes itself through specialized communication-efficient optimizers and hardware-aware acceleration techniques. By utilizing gradient compression, quantization
The framework distributes model parameters across multiple process groups by specifying the number of experts and the degree of expert parallelism.
Pythonbillion-parameterscompressiondata-parallelism
