b313a4867c
On Windows ROCm, torch.distributed ships without process-group helpers (is_initialized, is_available, get_rank, get_world_size). The worker subprocess already patches these in section 1e, but the main server process calls _determine_attention_impl_for_gpu_estimate() which calls unsloth's resolve_attention_implementation() → is_initialized(), causing: "Could not resolve attention implementation for '...': module 'torch.distributed' has no attribute 'is_initialized'" Fix: patch the missing attrs onto torch.distributed at the top of _determine_attention_impl_for_gpu_estimate, matching the same stubs already applied in worker.py section 1e. No-ops on Linux/CUDA where torch.distributed is fully populated. |
||
|---|---|---|
| .. | ||
| datasets | ||
| hardware | ||
| inference | ||
| models | ||
| paths | ||
| .gitkeep | ||
| __init__.py | ||
| _studio_release_build.py | ||
| cache_cleanup.py | ||
| downsample.py | ||
| native_path_leases.py | ||
| studio_version.py | ||
| subprocess_compat.py | ||
| transformers_version.py | ||
| update_status.py | ||
| utils.py | ||
| wheel_utils.py | ||