NVIDIA / NVIDIA/TensorRT-Model-Connect

[Community] Add a Hugging Face badge for verified Model Connect support

Open
#1,103 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Community
Dominant language
Python
Stars
254
Forks
58
Avg merge
1d 7h
Merged PRs (30d)
235

Description

Summary

Define a trustworthy Hugging Face badge or equivalent indicator for checkpoints and artifacts that have been verified with TensorRT Model Connect.

The indicator must represent qualification of an exact checkpoint and configuration. It must not imply that an untested checkpoint is fully supported merely because it belongs to a compatible model family.

Tasks

  • Define the precise meaning, scope, and qualification requirements of the indicator.
  • Require every support claim to identify:
    • Exact model or artifact ID.
    • Immutable checkpoint revision.
    • Precision and quantization configuration.
    • Target GPU and operating system.
    • TensorRT Model Connect, TensorRT, CUDA, and relevant runtime versions.
    • Qualification date and reproducible evidence.
  • Define how expired, superseded, or withdrawn qualifications are represented.
  • Determine the Hugging Face-supported integration mechanism for displaying the indicator.
  • Avoid repurposing library_name on third-party base-model repositories.
  • Do not modify third-party model cards without the repository owner's agreement.
  • Define an appropriate mechanism for NVIDIA-controlled artifacts and qualification records.
  • Generate badge or indicator metadata from qualification results where possible.
  • Link each displayed claim to reproducible commands, inputs, expected outputs, and qualification results.
  • Add validation that prevents family-level compatibility from being published as checkpoint-level verification.
  • Publish documentation explaining:
    • Verified support.
    • Best-effort compatibility.
    • Unsupported or untested configurations.
    • How qualification can be reproduced and renewed.
  • Establish ownership for approving, updating, and withdrawing indicators.

Acceptance criteria

At least one verified artifact displays the indicator through a Hugging Face-supported mechanism.

A user can trace the claim to reproducible qualification evidence for the exact checkpoint revision, precision, platform, and runtime versions.

The published indicator and documentation do not imply support for untested checkpoints or configurations in the same model family.

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Start by identifying any existing qualification, artifact, and documentation entry points in the repository, then confirm Hugging Face's supported mechanism for displaying a verified indicator. Define the exact-checkpoint evidence, lifecycle, ownership, validation, and documentation requirements, and finish when one artifact exposes a traceable claim without implying family-wide support.

Written by the indexing model from the issue text.

Assessment

Tech stack
huggingface
Domain
ai, documentation
Issue type
Feature
Difficulty
5/5
Estimated time
Over a week
Activity status
Active
Clarity
Needs clarification
Newbie friendliness
25/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.