Skip to content

Conversation

@abukhoy
Copy link
Contributor

@abukhoy abukhoy commented Nov 17, 2025

This feature adds support for exporting a proxy model, which disables the Embedding Layer and LM Head of a causal model.

Set enable_proxy = True to export the proxy model.
Set write_io = True to save input/output files during the generation stage.

Refer to the example script for implementation details.

Testing

Will update soon.

@vbaddi
Copy link
Contributor

vbaddi commented Nov 19, 2025

Thanks @abukhoy
Can we extend it to VLM also? Assuming Vision can have full model, the decoder_wrapper can have enable_proxy flag?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants