mirror of
https://github.com/saymrwulf/pytorch.git
synced 2026-05-14 20:57:59 +00:00
c4d835fbab
Fixes #142058 ## Summary DTensor `convolution_backward` op throws exception when the input Tensor has `requires_grad=False` which happens if the conv layer is the first layer in the model. ATEN convolution_backward op Usually returns 3 Tensors (grad_input, grad_weight, grad_bias) and the `grad_input` is actually an Optional[Tensor] which can be `None` in the case mentioned above. However, the DTensor sharding propagation rule and corresponding TP conv backward implementation both assume that the `grad_input` would be existent. ## Fix allow the `grad_input` to be `None` for `convolution_backward` op. ## Test `pytest test/distributed/tensor/test_convolution_ops.py` ## Follow-up The current implementation of DTensor conv op also ignores `output_mask` and this may need further care. Pull Request resolved: https://github.com/pytorch/pytorch/pull/142278 Approved by: https://github.com/bdhirsh |
||
|---|---|---|
| .. | ||
| _composable | ||
| _shard | ||
| _sharded_tensor | ||
| _sharding_spec | ||
| _symmetric_memory | ||
| _tensor | ||
| _tools | ||
| algorithms | ||
| autograd | ||
| benchmarks | ||
| checkpoint | ||
| elastic | ||
| examples | ||
| fsdp | ||
| launcher | ||
| nn | ||
| optim | ||
| pipelining | ||
| rpc | ||
| tensor | ||
| __init__.py | ||
| _checkpointable.py | ||
| _composable_state.py | ||
| _functional_collectives.py | ||
| _functional_collectives_impl.py | ||
| _serialization.py | ||
| _state_dict_utils.py | ||
| argparse_util.py | ||
| c10d_logger.py | ||
| collective_utils.py | ||
| constants.py | ||
| CONTRIBUTING.md | ||
| device_mesh.py | ||
| distributed_c10d.py | ||
| launch.py | ||
| logging_handlers.py | ||
| remote_device.py | ||
| rendezvous.py | ||
| run.py | ||
| utils.py | ||