[qwen-vl] fix position ids

[zucchini-nlp]

What does this PR do?

Fixes #40136 and fixes #40154

The accuracy on lm eval is restored. The issue was in position ids which weren't prepared correctly when generating. Using packed or unpacked positions doesn't affect anything, as I thought at first

Ig the values in position ids do not affect much generation results, given that slow tests weren't failing. We could add a test with the specific image/prompt where the outputs are significantly different, but I don't think it is necessary

[zucchini-nlp]

run-slow: qwen2_vl, qwen2_5_vl

[github-actions]

This comment contains run-slow, running the specified jobs:

models: ['models/qwen2_5_vl', 'models/qwen2_vl']
quantizations: [] ...

[HuggingFaceDocBuilderDev]

The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.

[zucchini-nlp]

run-slow: qwen2_5_vl, qwen2_vl

[github-actions]

This comment contains run-slow, running the specified jobs:

models: ['models/qwen2_5_vl', 'models/qwen2_vl']
quantizations: [] ...

[zucchini-nlp]

prob cc @vasqu for FA2 and packed attention

[zucchini-nlp]

run-slow: qwen2_5_omni, qwen2_5_vl, qwen2_vl

[github-actions]

This comment contains run-slow, running the specified jobs:

models: ['models/qwen2_5_omni', 'models/qwen2_5_vl', 'models/qwen2_vl']
quantizations: [] ...

[vasqu]

Just nits and question for my understanding

We could add a test with the specific image/prompt where the outputs are significantly different, but I don't think it is necessary

If it's not too much of a hassle, I'd appreciate it but no worries if not

[github-actions]

[For maintainers] Suggested jobs to run (before merge)

run-slow: qwen2_5_omni, qwen2_5_vl, qwen2_vl