Welcome to MoE Inference’s documentation!

Benchmarking

• Benchmarking
  • Evaluation plan
    • Orca approach
    • Proposed approach
  • Existing frameworks
    • Fairseq MoE
    • DeepSpeed-MoE
    • FasterTransformer MoE
  • MoE architectures
    • G-Shard
    • Fairseq’s 15B-parameters LM MoE model

Design and Implementation

• Design and Implementation
  • Legion Overview
    • Tasks
    • Regions
    • Partitioning
    • Control Replication
    • Coherence
    • Mapping
  • Design
  • Implementation

Indices and tables

• Index

• Module Index

• Search Page