u7-u7 c8f438c10a fix(environ): reject unsupported DEEPINFER_ACCELERATOR values ... The variable picks the deepinfer/_so/<accelerator>-<arch> directory that holds the extension module. The build scripts already reject anything but nvidia/ascend (build.sh, scripts/build_vendor.sh), but the runtime accepted any string: a typo such as "ascned" was silently coerced to the nvidia platform tag in _so_platform_tag(), so the process looked for a directory that does not exist and failed with a bare ModuleNotFoundError for deepinfer._C that never mentions the variable. Validate the value with EnvVar.choice so an unsupported value fails fast at import time with the variable name. The documented behaviours are unchanged: unset -> nvidia, "nvidia"/"ascend" -> themselves, and the empty value still means "let deep_cluster detect the accelerator".		2026-10-26 16:41:59 +08:00
..
distributed	initial commit	2026-10-20 12:09:42 +08:00
engine	initial commit	2026-10-20 12:09:42 +08:00
models	initial commit	2026-10-20 12:09:42 +08:00
ve	initial commit	2026-10-20 12:09:42 +08:00
worker	initial commit	2026-10-20 12:09:42 +08:00
__init__.py	initial commit	2026-10-20 12:09:42 +08:00
conftest.py	initial commit	2026-10-20 12:09:42 +08:00
test_cuda_sync_guards.py	initial commit	2026-10-20 12:09:42 +08:00
test_environ_accelerator.py	fix(environ): reject unsupported DEEPINFER_ACCELERATOR values	2026-10-26 16:41:59 +08:00