Skip to content

Feature: support Gemma 3n E4B (Gemma3nForConditionalGeneration) #2364

Open

Description

@qikail-ctrl

Request type

New model or checkpoint support

Problem and use case

Gemma 3n cannot be built or run with TensorRT-Model-Connect: no family handles Gemma3nForConditionalGeneration checkpoints (model_type: gemma3n), and the gemma family refuses them.

Proposed outcome

Build TensorRT engines for the checkpoint below and run images_text_to_text through the CLI and the C/C++ SDK, matching the Transformers reference implementation.

Model and target details

Alternatives considered

No response

Additional context

No response

Submission checks

  • I searched open and closed issues and found no duplicate request.
  • I removed secrets, private/internal evidence, personal paths, and restricted artifacts.

Activity

  1. self-assigned this
    on Nov 12, 2026
  2. qikail-ctrl commented on Nov 12, 2026

    @qikail-ctrl
    CollaboratorAuthor

    Pull request: #2365

  3. changed the title [-]Feature: support Gemma 3n (Gemma3nForConditionalGeneration)[/-] [+]Feature: support Gemma 3n E4B (Gemma3nForConditionalGeneration)[/+] on Nov 19, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

Labels

enhancementNew feature or request

Type

No type

Projects

No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions