LPLB: linear-programming load balancer for MoE expert parallelism (#24515)

Co-authored-by: xutizhou <xutingz@nvidia.com>
2026-06-16 10:19:42 -07:00
parent 00081a00d5
commit 92b42c8d8a
19 changed files with 2324 additions and 14 deletions
@@ -40,6 +40,7 @@ dependencies = [
  "numpy",
  "nvidia-cutlass-dsl[cu13]==4.5.2",
  "nvidia-ml-py",
+  "nvidia-mathdx==25.6.0",
  "openai-harmony==0.0.4",
  "openai==2.6.1",
  "orjson",