[BE] Move NoParallel to torchtitan.distributed

NoParallel should not belong `expert_parallel`. This PR moves it to `torchtitan.distributed.__init__.py`.

Author: fegin

torchtitan/distributed/__init__.py

@@ -5,7 +5,64 @@
 # LICENSE file in the root directory of this source tree.
 
 
+from torch.distributed.tensor import DeviceMesh, distribute_module, DTensor, Replicate
+from torch.distributed.tensor.parallel import ParallelStyle
+from torch.distributed.tensor.placement_types import Placement
+
 from torchtitan.distributed.parallel_dims import ParallelDims
 
 
-__all__ = ["ParallelDims"]
+__all__ = ["ParallelDims", "NoParallel"]
+
+
+# NOTE: This is to achieve replicate computation on the gate module in the MoE router.
+# It does nothing other than (1) setting the module parameters as DTensors on the given mesh
+# and (2) inserting hooks to module boundary to change torch.Tensor to DTensor and back.
+# The reason we need this wrapping is to ensure all parameters are on the same 1D/2D mesh,
+# which is assumed by (1) gradient norm clipping, and (2) optimizer fused implementation.
+class NoParallel(ParallelStyle):
+    def __init__(
+        self,
+        *,
+        input_layout: Placement | None = None,
+        output_layout: Placement | None = None,
+        use_local_output: bool = True,
+    ):
+        super().__init__()
+        self.input_layout = input_layout or Replicate()
+        self.output_layout = output_layout or Replicate()
+        self.desired_input_layout = Replicate()
+        self.use_local_output = use_local_output
+
+    @staticmethod
+    def _prepare_input_fn(input_layout, desired_input_layout, mod, inputs, device_mesh):
+        # annotate module input placements/sharding with input_layouts
+        input_tensor = inputs[0]
+        if not isinstance(input_tensor, DTensor):
+            input_tensor = DTensor.from_local(
+                input_tensor, device_mesh, (input_layout,), run_check=False
+            )
+
+        if input_layout != desired_input_layout:
+            input_tensor = input_tensor.redistribute(
+                placements=(desired_input_layout,), async_op=True
+            )
+        return (input_tensor, *inputs[1:])
+
+    @staticmethod
+    def _prepare_output_fn(output_layout, use_local_output, mod, outputs, device_mesh):
+        if outputs.placements != (output_layout,):
+            outputs = outputs.redistribute(placements=(output_layout,), async_op=True)
+        # back to local tensor
+        return outputs.to_local() if use_local_output else outputs
+
+    def _apply(self, module: nn.Module, device_mesh: DeviceMesh) -> nn.Module:
+        return distribute_module(
+            module,
+            device_mesh,
+            None,
+            partial(
+                self._prepare_input_fn, self.input_layout, self.desired_input_layout
+            ),
+            partial(self._prepare_output_fn, self.output_layout, self.use_local_output),
+        )

torchtitan/distributed/expert_parallel.py

@@ -5,7 +5,6 @@
 # LICENSE file in the root directory of this source tree.
 
 
-from functools import partial
 from typing import Callable, Literal
 
 import torch
@@ -16,11 +15,9 @@
     distribute_module,
     distribute_tensor,
     DTensor,
-    Replicate,
     Shard,
 )
 from torch.distributed.tensor.parallel import ParallelStyle
-from torch.distributed.tensor.placement_types import Placement
 
 
 # from torch.distributed._functional_collectives import all_to_all_single_autograd
@@ -108,59 +105,6 @@ def _apply(self, module: nn.Module, device_mesh: DeviceMesh) -> nn.Module:
         )
 
 
-# NOTE: This is to achieve replicate computation on the gate module in the MoE router.
-# It does nothing other than (1) setting the module parameters as DTensors on the given mesh
-# and (2) inserting hooks to module boundary to change torch.Tensor to DTensor and back.
-# The reason we need this wrapping is to ensure all parameters are on the same 1D/2D mesh,
-# which is assumed by (1) gradient norm clipping, and (2) optimizer fused implementation.
-class NoParallel(ParallelStyle):
-    def __init__(
-        self,
-        *,
-        input_layout: Placement | None = None,
-        output_layout: Placement | None = None,
-        use_local_output: bool = True,
-    ):
-        super().__init__()
-        self.input_layout = input_layout or Replicate()
-        self.output_layout = output_layout or Replicate()
-        self.desired_input_layout = Replicate()
-        self.use_local_output = use_local_output
-
-    @staticmethod
-    def _prepare_input_fn(input_layout, desired_input_layout, mod, inputs, device_mesh):
-        # annotate module input placements/sharding with input_layouts
-        input_tensor = inputs[0]
-        if not isinstance(input_tensor, DTensor):
-            input_tensor = DTensor.from_local(
-                input_tensor, device_mesh, (input_layout,), run_check=False
-            )
-
-        if input_layout != desired_input_layout:
-            input_tensor = input_tensor.redistribute(
-                placements=(desired_input_layout,), async_op=True
-            )
-        return (input_tensor, *inputs[1:])
-
-    @staticmethod
-    def _prepare_output_fn(output_layout, use_local_output, mod, outputs, device_mesh):
-        if outputs.placements != (output_layout,):
-            outputs = outputs.redistribute(placements=(output_layout,), async_op=True)
-        # back to local tensor
-        return outputs.to_local() if use_local_output else outputs
-
-    def _apply(self, module: nn.Module, device_mesh: DeviceMesh) -> nn.Module:
-        return distribute_module(
-            module,
-            device_mesh,
-            None,
-            partial(
-                self._prepare_input_fn, self.input_layout, self.desired_input_layout
-            ),
-            partial(self._prepare_output_fn, self.output_layout, self.use_local_output),
-        )
-
-
 class ExpertParallel(ParallelStyle):
     def __init__(self):
         super().__init__()

torchtitan/experiments/llama4/infra/parallelize.py

@@ -19,12 +19,11 @@
     SequenceParallel,
 )
 from torchtitan.config import JobConfig, TORCH_DTYPE_MAP
-from torchtitan.distributed import ParallelDims
+from torchtitan.distributed import NoParallel, ParallelDims
 
 from torchtitan.distributed.expert_parallel import (
     ExpertParallel,
     ExpertTensorParallel,
-    NoParallel,
     ReordererSequenceParallel,
     TensorParallel,
 )

torchtitan/experiments/qwen3/infra/parallelize.py

@@ -21,8 +21,7 @@
 )
 
 from torchtitan.config import JobConfig, TORCH_DTYPE_MAP
-from torchtitan.distributed import ParallelDims
-from torchtitan.distributed.expert_parallel import NoParallel
+from torchtitan.distributed import NoParallel, ParallelDims
 from torchtitan.models.llama3.infra.parallelize import (
     apply_ac,
     apply_compile,

torchtitan/models/deepseek_v3/infra/parallelize.py

@@ -17,8 +17,7 @@
 )
 
 from torchtitan.config import JobConfig, TORCH_DTYPE_MAP
-from torchtitan.distributed import ParallelDims
-from torchtitan.distributed.expert_parallel import NoParallel
+from torchtitan.distributed import NoParallel, ParallelDims
 from torchtitan.distributed.tensor_parallel import maybe_enable_async_tp
 from torchtitan.experiments.llama4.infra.parallelize import (
     apply_compile,