Repository navigation

laya: add native CUDA inference and optimize batch-1 latency - #90

Merged

hsliuustc0106 merged 5 commits into

InferMatrixLab:mainfrom

linear3735:codex/laya-batch1-cuda

Oct 5, 2026

Merged

laya: add native CUDA inference and optimize batch-1 latency#90
hsliuustc0106 merged 5 commits into
InferMatrixLab:mainfrom
linear3735:codex/laya-batch1-cuda

Commits on Oct 5, 2026
- Add the complete Laya CUDA worker and batch-1 kernel optimizations
  linear3735
  committed
  10c14a3
- Document the implemented Laya native worker boundaries
  linear3735
  committed
  eea6d76
- Remove blank context whitespace from the reproduction patch
  linear3735
  committed
  90f2336
- Record fresh Laya CUDA CLI and HTTP acceptance
  linear3735
  committed
  7b87adc
- Merge latest main documentation into Laya CUDA branch
  linear3735
  committed
  c2792b1