pytorch
a88bfc60 - [2/N][ST deprecate][BE] Remove Replicate Tensor convert from DDP and PTD (#95450)

Comment changes are shownComment changes are hidden
Commit
1 year ago
[2/N][ST deprecate][BE] Remove Replicate Tensor convert from DDP and PTD (#95450) No use is found for this ST/Replicated Tensor based DDP. As part of ShardedTensor migration, let's remove this logic. Trying to undo everything in https://github.com/pytorch/pytorch/pull/75753. Pull Request resolved: https://github.com/pytorch/pytorch/pull/95450 Approved by: https://github.com/wanchaol
Author
Committer
Parents
  • test/distributed
    • _shard
      • File
        test_replicated_tensor.py
    • optim
      • File
        test_zero_redundancy_optimizer.py
    • File
      test_c10d_gloo.py
  • torch
    • nn/parallel
      • File
        _replicated_tensor_ddp_interop.py
      • File
        _replicated_tensor_ddp_utils.py
      • File
        distributed.py
    • testing/_internal
      • File
        common_distributed.py
      • distributed
        • File
          distributed_test.py