Fix three messages that pass extra positional arguments

[ErenAta16]

What does this PR do?

Three messages pass an extra positional argument where a single formatted string was intended.

1 & 2 — VaeImageProcessor.__init__ and WanImageProcessor.__init__

raise ValueError(
    "`do_convert_rgb` and `do_convert_grayscale` can not both be set to `True`,"
    " if you intended to convert the image into RGB format, please set `do_convert_grayscale = False`.",
    " if you intended to convert the image into grayscale format, please set `do_convert_rgb = False`",
)

The last clause is a second argument, so Python puts each into args and str(e) renders the tuple repr:

('`do_convert_rgb` and `do_convert_grayscale` can not both be set to `True`, if you intended to convert the image into RGB format, please set `do_convert_grayscale = False`.', ' if you intended to convert the image into grayscale format, please set `do_convert_rgb = False`')

Half the guidance arrives wrapped in quotes and parentheses. That message is the only thing telling the caller which of the two flags to turn off, so both halves matter.

3 — pipeline_spectrogram_diffusion

logger.info("Generated segment", i)

logging treats i as a %-format argument for a message with no placeholder, formatting raises inside the handler, and logging catches it and writes --- Logging error --- plus a traceback to stderr. The per-segment progress line never reaches a handler — and this one is inside the denoising loop, so it fires once per segment.

Concatenated the two message halves; added %d to the log format.

An AST sweep over src/diffusers/ for multi-argument string raises and for logging calls with a placeholder-free first argument plus extra positional args reports these three. The fourth hit, examples/community/llm_grounded_diffusion.py:636, is in examples/ — happy to include it if you want examples in scope, left out to keep this to src/.

Before submitting

• Did you read the contributor guideline?
• Did you write any new necessary tests?

test_rgb_and_grayscale_conflict_message_is_one_string in tests/others/test_image_processor.py asserts len(e.args) == 1, that both halves of the guidance are present in str(e), and that the message doesn't start with (.

Stashing only image_processor.py gives AssertionError: assert 2 == 1; with the change, tests/others/test_image_processor.py is 11 passed. ruff check and ruff format --check clean on all four files.

Who can review?

@yiyixuxu @sayakpaul @DN6

[Solaris-star]

Reviewed — correct fix, well-scoped. Three distinct issues, all handled properly:

1. ValueError with multiple positional args (both image_processor.py and wan/image_processor.py): ValueError("a", "b", "c") creates a tuple of args, so str(e) renders ('a', 'b', 'c') instead of a clean concatenated message. Removing the commas between adjacent string literals lets Python implicitly concatenate them into one string. The test asserting len(ctx.exception.args) == 1 and not message.startswith("(") pins this nicely.

2. logger.info("Generated segment", i) in the spectrogram pipeline: the format string had no %d placeholder, so i was either silently dropped or would trigger a formatting error depending on the logging backend. Adding %d makes it proper lazy formatting.

One minor, non-blocking note: the spectrogram pipeline lives under deprecated/, so the logger fix is valid but low-impact. Not a blocker — fixing it everywhere is the right call for consistency.

The test coverage is solid for the ValueError case. Nice catch on the tuple-args rendering issue.

[github-actions]

Hi @ErenAta16, thanks for the PR! It does not appear to link an issue it fixes. If this PR addresses an existing issue, please add a closing keyword (e.g. Fixes #1234) to the PR description so the issue is linked. See the contribution guide for more details. If this PR intentionally does not fix a tracked issue, a maintainer can add the no-issue-needed label to silence this reminder.