Skip to content
Closed
Show file tree
Hide file tree
Changes from all commits
Commits
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
13 changes: 5 additions & 8 deletions axlearn/common/checkpointer_orbax_emergency.py
Original file line number Diff line number Diff line change
Expand Up @@ -751,7 +751,7 @@ def save(
# including step time in total blocking time.
start_t = time.perf_counter()
self._get_tensor_manager(state_with_tensors).save(
step=step, args=ocp.args.Composite(state=ocp.args.PyTreeSave(item=state_with_tensors))
step=step, args=ocp.args.PyTreeSave(item=state_with_tensors)
)
time_diff = time.perf_counter() - start_t
if self._composite_save_policy(step=step, evaler_summaries=self._eval_summaries):
Expand Down Expand Up @@ -808,9 +808,7 @@ def restore(

restored_state_with_tensors = tensor_manager.restore(
step=step,
args=ocp.args.Composite(
state=ocp.args.PyTreeRestore(item=self._get_abstract_state(state_with_tensors))
),
args=ocp.args.PyTreeRestore(item=self._get_abstract_state(state_with_tensors)),
)
# Merge non-tensor and tensor states by replacing leaves of the non-tensor Pytree with the
# not-None leaves of the tensor Pytree.
Expand All @@ -828,8 +826,7 @@ def wait_until_finished(self):
self._non_tensor_manager.wait_until_finished()
self._tensor_manager.wait_until_finished()

def stop(self, *, has_exception: bool = False):
def stop(self):
"""See `BaseCheckpointer.stop` for details."""
self._non_tensor_manager.stop(has_exception=has_exception)
if self._tensor_manager:
self._tensor_manager.close()
self._non_tensor_manager.stop()
self._tensor_manager.close()
2 changes: 1 addition & 1 deletion pyproject.toml
Original file line number Diff line number Diff line change
Expand Up @@ -154,7 +154,7 @@ mmau = [
# Orbax checkpointing.
orbax = [
"humanize==4.10.0",
"orbax-checkpoint==0.11.15",
"orbax-checkpoint==0.11.1",
]
# Audio dependencies.
audio = [
Expand Down