Skip to content

Mangle root scope TF variable name during tf.saved_model export #5738

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
merged 2 commits into from
Oct 26, 2023
Merged

Mangle root scope TF variable name during tf.saved_model export #5738

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
a6b1a71
mangle tf root scope name
lsy323 Oct 26, 2023
9943009
format
lsy323 Oct 26, 2023
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
15 changes: 14 additions & 1 deletion torch_xla/tf_saved_model_integration.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -58,6 +58,19 @@ def _make_input_signatures(
shape=spec.shape, dtype=getattr(tf, spec.dtype), name=f'args_{i}')


def _mangle_tf_root_scope_name(name):
# TF has more restricted constrain on the variable names at root scope.
# Root scope name constrain: [A-Za-z0-9.][A-Za-z0-9_.\\-/]*
# Non-root scope name constrain: [A-Za-z0-9_.\\-/]*
# https://github.com/tensorflow/tensorflow/blob/51b601fa6bb7e801c0b6ae73c25580e40a8b5745/tensorflow/python/framework/ops.py#L3301-L3302
# The state_dict key doesn't have such constrain,
# the name need to be mangled when a root-scoped TF variable is created.
if name[0] in "._\\-/":
return 'k' + name
else:
return name


def save_stablehlo_graph_as_tf(
stablehlo_program: stablehlo.StableHLOGraphModule,
path: os.PathLike,
Expand All @@ -81,7 +94,7 @@ def save_stablehlo_graph_as_tf(
bundle = copy.deepcopy(stablehlo_program._bundle)
tfm = tf.Module()
bundle.state_dict = {
k: tf.Variable(v, trainable=False, name=k)
k: tf.Variable(v, trainable=False, name=_mangle_tf_root_scope_name(k))
for k, v in bundle.state_dict.items()
}
bundle.additional_constants = [
Expand Down