Enabled configurable auto Tensor Parallelism (TP) for the inference of diverse models#6553

Open

gyou2021 wants to merge 6 commits intomicrosoft:masterfrom gyou2021:configurable_autoTP

+19-24

Commits on Sep 18, 2024

Enabled Qwen2-MoE Tensor Parallism (TP) inference
gyou2021
committed
Enabled configurable auto Tensor Parallelism (TP) for inference of diverse models
gyou2021
committed

Commits on Oct 25, 2024

Merge branch 'master' into configurable_autoTP
loadams
authored

Commits on Jan 8, 2025

Merge branch 'master' into configurable_autoTP
loadams
authored

Commits on Jan 13, 2025

Merge branch 'master' into configurable_autoTP
loadams
authored

Commits on Jan 17, 2025

Merge branch 'microsoft:master' into configurable_autoTP
gyou2021
authored