Skip to content

Conversation

@quic-dhirajku
Copy link
Contributor

Added support for Qwen3ForCausalLM models, tested on Qwen3-0.6B model for CI runs. Updated modeling internvl script to allow proper prefix chunking of vision+embeds when more than 1 patches are needed. Test InternVL_3_5_1B model for 1 and full layers via CI.

Added support for Qwen3ForCausalLM models, tested on Qwen3-0.6B model for CI runs.
Updated modeling internvl script to allow proper prefix chunking of vision+embeds when more than 1 patches are needed.
Test InternVL_3_5_1B model for 1 and full layers via CI.

Signed-off-by: quic-dhirajku <[email protected]>
Updated internvl_inference script to allow easy batch inference and compilation.
This method supports single prompt single image batching method as originally supported by the model and in the same template.

Signed-off-by: quic-dhirajku <[email protected]>
Tested end to end runs for 1B version of 2.5,3,3.5 models.

Signed-off-by: quic-dhirajku <[email protected]>
@vbaddi
Copy link
Contributor

vbaddi commented Oct 14, 2025

@quic-dhirajku can your run the pre-commit install and resolve the lint/format issues?

Signed-off-by: quic-dhirajku <[email protected]>
@quic-hemagnih quic-hemagnih merged commit 1272415 into quic:main Oct 14, 2025
4 checks passed
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.

3 participants