Skip to content
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

Jump to bottom

Choose quantization modifier based on dynamic set of modifier properties #1568

Merged
merged 2 commits into from
May 23, 2023
Merged

Choose quantization modifier based on dynamic set of modifier properties #1568

Changes from 1 commit
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
4101329
Choose quantization modifier based on dynamic set of modifier properties
KSGulin May 18, 2023
ecfd614
Merge branch 'main' into update_mod_select
KSGulin May 23, 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
21 changes: 20 additions & 1 deletion src/sparseml/pytorch/sparsification/quantization/modifier_quantization.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -67,7 +67,26 @@
def _select_quantization_modifier(state: Dict[str, Any]) -> Type:
# if kwargs for the legacy quantization modifier are provided,
# route YAML loading to that class
return LegacyQuantizationModifier if "submodules" in state else QuantizationModifier
legacy_props = set(
[
str(prop)
for prop in dir(LegacyQuantizationModifier)
if isinstance(getattr(LegacyQuantizationModifier, prop), ModifierProp)
]
)
current_props = set(
[
str(prop)
for prop in dir(QuantizationModifier)
if isinstance(getattr(QuantizationModifier, prop), ModifierProp)
]
)
legacy_only_props = legacy_props - current_props
return (
LegacyQuantizationModifier
if any(field in state for field in legacy_only_props)
KSGulin marked this conversation as resolved.
Show resolved Hide resolved
else QuantizationModifier
)


@PyTorchModifierYAML(swap_class_by_state_fn=_select_quantization_modifier)
Expand Down
Loading