Enable native AMD ROCm inference via Triton kernels#166

Merged

jandom merged 5 commits intoaqlaboratory:mainfrom

singagan:amd-rocm-inference

Apr 9, 2026

Commits on Apr 5, 2026

Enable native AMD ROCm inference via Triton kernels
singagan
committed

Commits on Apr 7, 2026

Add AMD ROCm documentation and environment validation
singagan
committed

Commits on Apr 8, 2026

Remove oudated pae_enabled model setting from triton.yml
jnwei
authored
Fix test failures on ROCm: skip DeepSpeed tests, use Triton in model tests
singagan
committed

Commits on Apr 9, 2026

Fix HIPBLAS_STATUS_INTERNAL_ERROR on ROCm by setting preferred BLAS library in test conftest
singagan
committed