viable/strict/1789588240: Move the miopen_batch_norm decomposition to the core table (#197160)
- PyTorch: 1081 events in the last 90 days
- PyTorch: 1068th Release in the last 90 days
- Previous: earlier the same day · viable/strict/1789586681
What happened
DTensor's sharding propagator has no strategy for aten.miopen_batch_norm , so F.batch_norm on a sharded 4D input fails on ROCm with "Operator aten.miopen_batch_norm.default does not have a sharding strategy registered". CUDA never hits the equivalent because aten.cudnn_batch_norm has a decomposition in torch/_decomp , which DTensor's DecompShardingStrategy falls back to. The ROCm decomposition existed only in torch/…
Summary assembled by rule from the sources below