4d09cbbda4
torchao.float8.inference accesses ProcessGroup.BackendType as a class-level attribute. Plain type() stubs have no __getattr__ on the metaclass so this raises AttributeError. Introduce _StubClassMeta whose __getattr__ returns child stub classes, fixing the torchao import chain. Add an amd-smi circuit breaker in amd.py: after 3 consecutive failures the module stops spawning the process, eliminating the repeated Windows UAC / DiskPart elevation prompts caused by polling a non-functional amd-smi. Also guard BNB_ROCM_VERSION=72 behind a DLL existence check so bitsandbytes fails with its own detection message rather than a harder "DLL not found" when the Windows ROCm bnb wheel is not yet installed. |
||
|---|---|---|
| .. | ||
| __init__.py | ||
| amd.py | ||
| hardware.py | ||
| nvidia.py | ||
| VRAM_ESTIMATION.md | ||
| vram_estimation.py | ||