You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Add two CUDA 12.6 matrix entries to the release workflow, matching
the second-state/qwen3_asr_rs release structure:
transcribe-linux-x86_64-cuda.zip — CUDA 12.6 libtorch (x86_64)
transcribe-linux-aarch64-cuda.zip — CUDA 12.6 libtorch (Jetson/Grace)
Each CUDA variant installs cuda-toolkit-12-6 for linking (aarch64 also
adds libopenblas-dev), sets -rpath-link and LD_LIBRARY_PATH so the
linker resolves CUDA symbols from libtorch, and bundles the CUDA-enabled
libtorch/lib/ in the release zip alongside both binaries and vocab.json.
Signed-off-by: Michael Yuan <michael@secondstate.io>
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
0 commit comments