[Bugfix] Fix GLM4.1V multimodal processor with compatability for Transformers v4.56

[Isotr0py]

Purpose

• Fix [VLM] Optimize GLM4.5-V-style video processing to only decode necessary frames #24161 (comment)
• Separate from [Model] Support Qwen3-VL Model Series #24727
• Add Transformer v4.56's video proceesing compatability for GLM4.1V

Test Plan

pytest -s -v tests/models/multimodal/processing/test_glm4_1v.py -k test_processor_override

Test Result

Test should pass.

---

Essential Elements of an Effective PR Description Checklist

• The purpose of the PR, such as "Fix some issue (link existing issues this PR will resolve)".
• The test plan, such as providing test command.
• The test results, such as pasting the results comparison before and after, or e2e results
• (Optional) The necessary documentation update, such as updating supported_models.md and examples for a new model.
• (Optional) Release notes update. If your change is user facing, please update the release notes draft in the Google Doc.

[DarkLight1337]

Let's see if transformers nightly passes now

[DarkLight1337]

Well, Intern-S1 is still failing, so can't check it

[DarkLight1337]

Oh, this breaks the existing processor test

[Isotr0py]

Well, Intern-S1 is still failing, so can't check it

I see, let me check it with nightly Transformers locally then.

[Isotr0py]

Have confirmed v4.56.1 and v4.57.0.dev(huggingface/transformers@9c804f7) can still pass the GLM4.1V tests in tests/models/multimodal/processing/test_glm4_1v.py and tests/models/multimodal/processing/test_common.py