Skip to content

[OV] Fix VLM in-place static quantization #1464

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 1 commit into from
Oct 8, 2025
Merged

[OV] Fix VLM in-place static quantization #1464

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
9 changes: 9 additions & 0 deletions optimum/intel/openvino/quantization.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1440,6 +1440,15 @@ def _quantize_ovbasemodel(
quantized_model = _mixed_quantization(submodel, config, nncf_dataset, **kwargs)

# Replace the original model with the quantized model
if isinstance(self.model, OVModelForVisualCausalLM):
# Special handling of submodels in OVModelForVisualCausalLM
# TODO (nikita.savelyevv): Implement a proper fix including other model types
if submodel_name == "lm_model":
self.model.language_model.model = quantized_model
elif submodel_name == "text_embeddings_model":
self.model.language_model.text_emb_model = quantized_model
elif submodel_name == "vision_embeddings_model":
self.model.vision_embeddings.model = quantized_model
if isinstance(getattr(self.model, submodel_name), openvino.Model):
setattr(self.model, submodel_name, quantized_model)
elif isinstance(getattr(getattr(self.model, submodel_name), "model"), openvino.Model):
Expand Down
Loading