[TRTLLM-13378][feat] Drop legacy --extra_visual_gen_options CLI alias by zhenhuaw-me · Pull Request #15262 · NVIDIA/TensorRT-LLM

zhenhuaw-me · 2026-06-11T11:48:32Z

Description

The standard --visual_gen_args flag has stabilized and all in-tree examples, docs, and benchmark scripts use it. This drops the legacy --extra_visual_gen_options click/argparse alias (kept until now as an alias of --visual_gen_args in trtllm-serve, trtllm-bench visual-gen, the standalone visual-gen benchmark script, and the offline example scripts) and updates the remaining doc, README, YAML-comment, and shell-script references to point at the standard flag.

CLI definitions: tensorrt_llm/commands/serve.py, tensorrt_llm/bench/benchmark/visual_gen.py, tensorrt_llm/serve/scripts/benchmark_visual_gen.py
Docs: docs/source/commands/trtllm-serve/trtllm-serve.rst
Examples: examples/visual_gen/serve/README.md, both benchmark_visual_gen*.sh, comment headers in examples/visual_gen/configs/*.yaml, and the argparse aliases in examples/visual_gen/models/*.py
Eval: scripts/visualgen_eval/visual_gen_lpips_score_eval.py

After this change, git grep extra_visual_gen_options returns nothing.

Test Coverage

Built TRT-LLM from source on a GB200 dev container and confirmed:

trtllm-serve serve --help lists only --visual_gen_args.
trtllm-bench visual-gen --help lists only --visual_gen_args.
--extra_visual_gen_options is absent from both.

Tracking: TRTLLM-13378 (under epic TRTLLM-10897 — VisualGen API Polish (M2)).

PR Checklist

Please check this after reviewing the above items as appropriate for this PR.

Summary by CodeRabbit

Documentation
- Updated visual generation serving documentation, example scripts, and benchmark tools. The configuration flag for visual generation has been renamed to --visual_gen_args across all examples and serving commands.

zhenhuaw-me · 2026-06-11T12:10:23Z

/bot run

coderabbitai · 2026-06-11T12:14:18Z

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Enterprise

Run ID: aec778a9-12f2-4bdb-8ee4-e3509554ee6b

📥 Commits

Reviewing files that changed from the base of the PR and between 835fd61 and cef8084.

📒 Files selected for processing (18)

docs/source/commands/trtllm-serve/trtllm-serve.rst
examples/visual_gen/configs/ltx2-4gpu.yaml
examples/visual_gen/configs/wan2.2-t2v-fp4-1gpu.yaml
examples/visual_gen/configs/wan2.2-t2v-fp4-4gpu.yaml
examples/visual_gen/configs/wan2.2-t2v-fp8-8gpu.yaml
examples/visual_gen/models/cosmos3_ti2v.py
examples/visual_gen/models/flux1.py
examples/visual_gen/models/flux2.py
examples/visual_gen/models/ltx2.py
examples/visual_gen/models/wan_i2v.py
examples/visual_gen/models/wan_t2v.py
examples/visual_gen/serve/README.md
examples/visual_gen/serve/benchmark_visual_gen.sh
examples/visual_gen/serve/benchmark_visual_gen_mgmn_distributed.sh
scripts/visualgen_eval/visual_gen_lpips_score_eval.py
tensorrt_llm/bench/benchmark/visual_gen.py
tensorrt_llm/commands/serve.py
tensorrt_llm/serve/scripts/benchmark_visual_gen.py

💤 Files with no reviewable changes (9)

examples/visual_gen/models/ltx2.py
examples/visual_gen/models/cosmos3_ti2v.py
examples/visual_gen/models/wan_i2v.py
tensorrt_llm/commands/serve.py
examples/visual_gen/models/flux2.py
examples/visual_gen/models/flux1.py
examples/visual_gen/models/wan_t2v.py
tensorrt_llm/serve/scripts/benchmark_visual_gen.py
tensorrt_llm/bench/benchmark/visual_gen.py

📝 Walkthrough

Walkthrough

The PR renames a CLI flag across the TensorRT-LLM visual generation suite from --extra_visual_gen_options to --visual_gen_args. Changes include documentation updates, example script rewiring, core library command definitions, and supporting benchmark/evaluation utilities.

Changes

Visual Generation CLI Flag Rename

Layer / File(s)	Summary
Documentation and configuration examples `docs/source/commands/trtllm-serve/trtllm-serve.rst`, `examples/visual_gen/configs/*.yaml`, `examples/visual_gen/serve/README.md`	User-facing documentation and YAML config headers updated to reference `--visual_gen_args` instead of `--extra_visual_gen_options`, with example `trtllm-serve` commands revised accordingly.
Example model CLI argument wiring `examples/visual_gen/models/cosmos3_ti2v.py`, `examples/visual_gen/models/flux1.py`, `examples/visual_gen/models/flux2.py`, `examples/visual_gen/models/ltx2.py`, `examples/visual_gen/models/wan_i2v.py`, `examples/visual_gen/models/wan_t2v.py`	Argparse configuration in example scripts updated to remove `--extra_visual_gen_options` option declarations, consolidating on `--visual_gen_args` with `dest="visual_gen_args"` and unchanged downstream YAML loading.
Core library serve and benchmark CLI `tensorrt_llm/commands/serve.py`, `tensorrt_llm/bench/benchmark/visual_gen.py`, `tensorrt_llm/serve/scripts/benchmark_visual_gen.py`	Click and argparse decorators updated to expose only `--visual_gen_args` (or `--visual-gen-args` in hyphenated form) and remove previous flag aliases, preserving YAML path parameter bindings.
Benchmark and evaluation script updates `examples/visual_gen/serve/benchmark_visual_gen.sh`, `examples/visual_gen/serve/benchmark_visual_gen_mgmn_distributed.sh`, `scripts/visualgen_eval/visual_gen_lpips_score_eval.py`	Bash scripts updated to append `--visual_gen_args` flag when constructing `trtllm-serve` commands; Python evaluation utility docstring and help text reference the new flag name.

Estimated code review effort

🎯 2 (Simple) | ⏱️ ~10 minutes

🚥 Pre-merge checks | ✅ 4 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Docstring Coverage	⚠️ Warning	Docstring coverage is 0.00% which is insufficient. The required threshold is 80.00%.	Write docstrings for the functions missing them to satisfy the coverage threshold.

✅ Passed checks (4 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title clearly and specifically describes the main change: removing the legacy CLI alias in favor of the standard flag.
Description check	✅ Passed	The description includes all required sections: what was changed, why, test coverage details, and a completed PR checklist.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing Touches

🧪 Generate unit tests (beta)

Create PR with unit tests

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

tensorrt-cicd · 2026-06-11T12:17:04Z

PR_Github #53565 [ run ] triggered by Bot. Commit: cef8084 Link to invocation

tensorrt-cicd · 2026-06-11T13:06:11Z

PR_Github #53565 [ run ] completed with state SUCCESS. Commit: cef8084
/LLM/main/L0_MergeRequest_PR pipeline #42714 completed with status: 'FAILURE'

CI Report

⚠️ Action Required:

Please check the failed tests and fix your PR
If you cannot view the failures, ask the CI triggerer to share details
Once fixed, request an NVIDIA team member to trigger CI again

CI Agent Failure Analysis

Link to invocation

kaiyux

Approving on tensorrt_llm/bench changes

The standard --visual_gen_args flag has stabilized and all in-tree examples, docs, and benchmark scripts now use it. Remove the legacy --extra_visual_gen_options click/argparse alias from trtllm-serve, trtllm-bench visual-gen, the standalone visual-gen benchmark script, and the offline example scripts. Update remaining doc, README, YAML comment, and shell-script references to point at the standard flag. Signed-off-by: Zhenhua Wang <zhenhuaw@nvidia.com>

zhenhuaw-me · 2026-06-12T03:32:12Z

/bot run

tensorrt-cicd · 2026-06-12T03:37:49Z

PR_Github #53773 [ run ] triggered by Bot. Commit: 9fe12e9 Link to invocation

tensorrt-cicd · 2026-06-12T06:42:42Z

PR_Github #53773 [ run ] completed with state FAILURE. Commit: 9fe12e9
/LLM/main/L0_MergeRequest_PR pipeline #42893 completed with status: 'FAILURE'

CI Report

⚠️ Action Required:

Please check the failed tests and fix your PR
If you cannot view the failures, ask the CI triggerer to share details
Once fixed, request an NVIDIA team member to trigger CI again

CI Agent Failure Analysis

Link to invocation

zhenhuaw-me · 2026-06-12T06:55:32Z

/bot run

tensorrt-cicd · 2026-06-12T07:00:55Z

PR_Github #53827 [ run ] triggered by Bot. Commit: 9fe12e9 Link to invocation

tensorrt-cicd · 2026-06-12T07:24:33Z

PR_Github #53827 [ run ] completed with state FAILURE. Commit: 9fe12e9
/LLM/main/L0_MergeRequest_PR pipeline #42942 completed with status: 'FAILURE'

CI Report

⚠️ Action Required:

Please check the failed tests and fix your PR
If you cannot view the failures, ask the CI triggerer to share details
Once fixed, request an NVIDIA team member to trigger CI again

CI Agent Failure Analysis

Link to invocation

zhenhuaw-me · 2026-06-12T08:26:39Z

/bot run

tensorrt-cicd · 2026-06-12T08:33:13Z

PR_Github #53848 [ run ] triggered by Bot. Commit: 9fe12e9 Link to invocation

tensorrt-cicd · 2026-06-12T11:14:20Z

PR_Github #53848 [ run ] completed with state SUCCESS. Commit: 9fe12e9
/LLM/main/L0_MergeRequest_PR pipeline #42957 completed with status: 'FAILURE'

CI Report

⚠️ Action Required:

Please check the failed tests and fix your PR
If you cannot view the failures, ask the CI triggerer to share details
Once fixed, request an NVIDIA team member to trigger CI again

CI Agent Failure Analysis

Link to invocation

zhenhuaw-me · 2026-06-12T12:43:59Z

/bot run --disable-fail-fast

tensorrt-cicd · 2026-06-12T12:49:55Z

PR_Github #53885 [ run ] triggered by Bot. Commit: 9fe12e9 Link to invocation

tensorrt-cicd · 2026-06-12T19:57:34Z

PR_Github #53885 [ run ] completed with state SUCCESS. Commit: 9fe12e9
/LLM/main/L0_MergeRequest_PR pipeline #42985 completed with status: 'FAILURE'

CI Report

⚠️ Action Required:

Please check the failed tests and fix your PR
If you cannot view the failures, ask the CI triggerer to share details
Once fixed, request an NVIDIA team member to trigger CI again

CI Agent Failure Analysis

Link to invocation

zhenhuaw-me · 2026-06-13T01:38:03Z

/bot skip --comment "this is a trivial change and the remaining failures are H100 unrelated"

tensorrt-cicd · 2026-06-13T01:43:31Z

PR_Github #53983 [ skip ] triggered by Bot. Commit: 9fe12e9 Link to invocation

tensorrt-cicd · 2026-06-13T01:49:27Z

PR_Github #53983 [ skip ] completed with state SUCCESS. Commit: 9fe12e9
Skipping testing for commit 9fe12e9

Link to invocation

…NVIDIA#15262) Signed-off-by: Zhenhua Wang <zhenhuaw@nvidia.com> Signed-off-by: GitLab CI Bot <gitlab-ci@nvidia.com>

github-actions Bot assigned zhenhuaw-me Jun 11, 2026

zhenhuaw-me marked this pull request as ready for review June 11, 2026 12:10

zhenhuaw-me requested review from a team as code owners June 11, 2026 12:10

zhenhuaw-me requested review from FrankD412, arysef, mikeiovine and nv-guomingz June 11, 2026 12:10

zhenhuaw-me requested review from QiJune and chang-l June 11, 2026 12:51

kaiyux approved these changes Jun 12, 2026

View reviewed changes

zhenhuaw-me force-pushed the drop-legacy-cli branch from cef8084 to 9fe12e9 Compare June 12, 2026 03:32

chang-l reviewed Jun 15, 2026

View reviewed changes

Comment thread tensorrt_llm/commands/serve.py

Comment thread tensorrt_llm/bench/benchmark/visual_gen.py

Comment thread tensorrt_llm/serve/scripts/benchmark_visual_gen.py

chang-l approved these changes Jun 16, 2026

View reviewed changes

zhenhuaw-me merged commit 0593968 into NVIDIA:main Jun 17, 2026
7 checks passed

zhenhuaw-me deleted the drop-legacy-cli branch June 17, 2026 06:14

Uh oh!

Conversation

zhenhuaw-me commented Jun 11, 2026 • edited by coderabbitai Bot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Description

Test Coverage

PR Checklist

Summary by CodeRabbit

Uh oh!

zhenhuaw-me commented Jun 11, 2026

Uh oh!

coderabbitai Bot commented Jun 11, 2026

Walkthrough

Changes

Estimated code review effort

❌ Failed checks (1 warning)

Uh oh!

tensorrt-cicd commented Jun 11, 2026

Uh oh!

tensorrt-cicd commented Jun 11, 2026

Uh oh!

kaiyux left a comment

Choose a reason for hiding this comment

Uh oh!

zhenhuaw-me commented Jun 12, 2026

Uh oh!

tensorrt-cicd commented Jun 12, 2026

Uh oh!

tensorrt-cicd commented Jun 12, 2026

Uh oh!

zhenhuaw-me commented Jun 12, 2026

Uh oh!

tensorrt-cicd commented Jun 12, 2026

Uh oh!

tensorrt-cicd commented Jun 12, 2026

Uh oh!

zhenhuaw-me commented Jun 12, 2026

Uh oh!

tensorrt-cicd commented Jun 12, 2026

Uh oh!

tensorrt-cicd commented Jun 12, 2026

Uh oh!

zhenhuaw-me commented Jun 12, 2026

Uh oh!

tensorrt-cicd commented Jun 12, 2026

Uh oh!

tensorrt-cicd commented Jun 12, 2026

Uh oh!

zhenhuaw-me commented Jun 13, 2026

Uh oh!

tensorrt-cicd commented Jun 13, 2026

Uh oh!

tensorrt-cicd commented Jun 13, 2026

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

zhenhuaw-me commented Jun 11, 2026 •

edited by coderabbitai Bot

Loading