Skip to content

Fix pyink formatting #4939

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
copybara-service merged 1 commit into from
Aug 19, 2026
+12 −38
Merged

Fix pyink formatting #4939

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
50 changes: 12 additions & 38 deletions src/maxtext/trainers/diloco/utils/spmd_diloco_checkpointing.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -46,13 +46,9 @@ def restore_diloco_checkpoint(
)
)
restore_args = ocp.checkpoint_utils.construct_restore_args(diloco_abstract)
restored = ocp.args.PyTreeRestore(
item=diloco_abstract, restore_args=restore_args, partial_restore=True
)
restored = ocp.args.PyTreeRestore(item=diloco_abstract, restore_args=restore_args, partial_restore=True)
restored = ckptr.restore(epath.Path(path), args=restored)
return from_diloco_checkpoint_dict(
restored, abstract_nnx_state, config=config
)
return from_diloco_checkpoint_dict(restored, abstract_nnx_state, config=config)


def is_diloco_checkpoint(restored_dict: Any) -> bool:
Expand Down Expand Up @@ -103,14 +99,9 @@ def _add_diloco_dim(leaf):
if hasattr(leaf, "shape") and hasattr(leaf, "dtype"):
new_shape = (num_replicas, *leaf.shape)
sharding = getattr(leaf, "sharding", None)
if (
isinstance(sharding, jax.sharding.NamedSharding)
and "diloco" in sharding.mesh.axis_names
):
if isinstance(sharding, jax.sharding.NamedSharding) and "diloco" in sharding.mesh.axis_names:
new_spec = jax.sharding.PartitionSpec("diloco", *sharding.spec)
sharding = jax.sharding.NamedSharding(
mesh=sharding.mesh, spec=new_spec
)
sharding = jax.sharding.NamedSharding(mesh=sharding.mesh, spec=new_spec)
if isinstance(leaf, jax.ShapeDtypeStruct):
return jax.ShapeDtypeStruct(new_shape, leaf.dtype, sharding=sharding)
return jnp.broadcast_to(leaf, new_shape)
Expand All @@ -120,9 +111,7 @@ def _add_diloco_dim(leaf):

if hasattr(state, "model"):
_, params, _ = nnx.split(state.model, nnx.Param, ...)
params = (
params.to_pure_dict() if hasattr(params, "to_pure_dict") else params
)
params = params.to_pure_dict() if hasattr(params, "to_pure_dict") else params
elif hasattr(state, "params"):
params = state.params
else:
Expand All @@ -134,9 +123,7 @@ def _add_diloco_dim(leaf):
nesterov=True,
)
outer_opt_state = outer_optimizer.init(params)
step = getattr(
getattr(state, "optimizer", None), "step", jnp.array(0, dtype=jnp.int32)
)
step = getattr(getattr(state, "optimizer", None), "step", jnp.array(0, dtype=jnp.int32))

# 1. Inner state: convert per-replica NNX state to Linen checkpoint layout
if isinstance(inner_state, (nnx.State, train_state_nnx.TrainStateNNX)):
Expand Down Expand Up @@ -210,10 +197,7 @@ def _add_diloco_dim(leaf):
if hasattr(leaf, "shape") and hasattr(leaf, "dtype"):
new_shape = (num_replicas, *leaf.shape)
sharding = getattr(leaf, "sharding", None)
if (
isinstance(sharding, jax.sharding.NamedSharding)
and "diloco" in sharding.mesh.axis_names
):
if isinstance(sharding, jax.sharding.NamedSharding) and "diloco" in sharding.mesh.axis_names:
new_spec = jax.sharding.PartitionSpec("diloco", *sharding.spec)
sharding = jax.sharding.NamedSharding(mesh=sharding.mesh, spec=new_spec)
if isinstance(leaf, jax.ShapeDtypeStruct):
Expand All @@ -227,24 +211,18 @@ def _add_diloco_dim(leaf):
if isinstance(abstract_diloco_state, diloco.DiLoCoTrainState):
abstract_inner = abstract_diloco_state.inner_state
else:
abstract_inner = jax.tree_util.tree_map(
_add_diloco_dim, abstract_diloco_state
)
abstract_inner = jax.tree_util.tree_map(_add_diloco_dim, abstract_diloco_state)

if isinstance(abstract_inner, nnx.Module):
abstract_inner = nnx.state(abstract_inner)

if isinstance(abstract_inner, (nnx.State, train_state_nnx.TrainStateNNX)):
linen_state, aux_state, ephemeral = train_state_nnx.split_for_checkpoint(
abstract_inner
)
linen_state, aux_state, ephemeral = train_state_nnx.split_for_checkpoint(abstract_inner)
weights = train_state_nnx.from_linen_checkpoint_dict(inner_dict)
if "model" in weights:
nnx.replace_by_pure_dict(linen_state, {"model": weights["model"]})
if "optimizer" in weights:
nnx.replace_by_pure_dict(
linen_state, {"optimizer": weights["optimizer"]}
)
nnx.replace_by_pure_dict(linen_state, {"optimizer": weights["optimizer"]})
nnx_aux = inner_dict.get("nnx_aux")
if nnx_aux:
nnx.replace_by_pure_dict(aux_state, nnx_aux)
Expand Down Expand Up @@ -283,13 +261,9 @@ def _add_diloco_dim(leaf):
if isinstance(abstract_diloco_state, diloco.DiLoCoTrainState):
abstract_inner = abstract_diloco_state.inner_state
else:
abstract_inner = jax.tree_util.tree_map(
_add_diloco_dim, abstract_diloco_state
)
abstract_inner = jax.tree_util.tree_map(_add_diloco_dim, abstract_diloco_state)

inner_state = replace_nnx_model_params(
abstract_inner, broadcasted_model_params
)
inner_state = replace_nnx_model_params(abstract_inner, broadcasted_model_params)

outer_optimizer = optax.sgd(
getattr(config, "diloco_outer_lr", 0.1),
Expand Down
Loading