fix: resolve integration test failures by migrating COCO image loading to HF Hub fixtures

[abhishekkapoorx]

What does this PR do?

This PR resolves the integration test failures triaged in #47205.

The domain images.cocodataset.org has been unstable, resulting in connection timeouts and failing integration tests across multiple model pipelines. To address this, we have migrated COCO-dataset image loading in our test suites to use the mirrored fixtures on the Hugging Face Hub under hf-internal-testing/fixtures-coco and utilizing local path redirection.

Key Changes:

1. Added test helper: Introduced load_test_image in tests/test_processing_common.py which wraps load_image and automatically processes remote URLs with url_to_local_path.
2. Refactored tests: Converted occurrences of direct requests.get or httpx.get targeting http://images.cocodataset.org/val2017/000000039769.jpg to the new load_test_image helper pointing to the Hugging Face mirror.
3. Cleaned up CI downloads: Removed direct references to stale images.cocodataset.org URLs from the URLS_FOR_TESTING_DATA registry in utils/fetch_hub_objects_for_ci.py.

This prevents flaky network issues and leverages local caching during CI execution.

Related Issues

Fixes #47205

Code Agent Policy

The Transformers repo is currently being overwhelmed by a large number of PRs and issue comments written by
code agents. These often are low-quality, or fix extremely minor issues that occur rarely or never in practice.
As a result, we're instituting a rule that first-time contributors should not use code agents to submit PRs or issues.
We'd also ask autonomous "OpenClaw"-like agents not to open any PRs or issues.

Issues/PRs from first-time contributors that violate this rule will probably just be closed without review, and we
might block you, especially if you open more than one or appear to be deliberately ignoring this. We especially do not
want new contributors to jump in on random issues to contribute an agent-written fix. This creates lots of noise
for reviewers and other users and will almost certainly get you blocked.

For more information, please read CONTRIBUTING.md.

• (First-time contributors only): I confirm that this PR description and code is not written by an LLM or code agent

Before submitting

• This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case).
• Did you read the contributor guideline and the
  Pull Request checks?
• Was this discussed/approved via a Github issue or the forum? Please add a link
  to it if that's the case.
• Did you make sure to update the documentation with your changes according to the guidelines?
• Did you write any new necessary tests?

Who can review?

Anyone in the community is free to review the PR once the tests have passed. Feel free to tag
members/contributors who may be interested in your PR.

[github-actions]

---

CI recap

Dashboard: View test results in Grafana
Latest run: 29143766271:2
Result: failure | Jobs: 2 | Tests: 21 | Failures: 1 | Duration: 2m 18s

Code quality check failed: test jobs were skipped. Fix the code quality issues and push again to run tests.

[github-actions]

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

run-slow: aimv2, aria, deepseek_vl_hybrid, eomt, eomt_dinov3, fast_vlm, flava, fuyu, idefics, imagegpt, kimi_k25, layoutlmv2, lfm2_vl, llava, minimax_m3_vl, pix2struct

[abhishekkapoorx]

@ydshieh In src/transformers/modeling_utils.py, we are using the axis but latest documentation of torch uses dim. I think we should fix this as well.

index 3f14d74979..f27ae1d884 100644
--- a/src/transformers/modeling_utils.py
+++ b/src/transformers/modeling_utils.py
@@ -982,7 +982,7 @@ class ModuleUtilsMixin:
                     torch.ones((batch_size, seq_length, prefix_seq_len), device=device, dtype=causal_mask.dtype),
                     causal_mask,
                 ],
-                axis=-1,
+                dim=-1,
             )
 
         extended_attention_mask = causal_mask[:, None, :, :] * attention_mask[:, None, None, :]

This is causing local typing check to fail.

[tarekziade]

run-slow: aimv2, aria, deepseek_vl_hybrid, eomt, eomt_dinov3, fast_vlm, flava, fuyu, idefics, imagegpt, kimi_k25, layoutlmv2, lfm2_vl, llava, minimax_m3_vl, pix2struct

[github-actions]

Workflow Run ⚙️

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

models: ["models/aimv2", "models/aria", "models/deepseek_vl_hybrid", "models/eomt", "models/eomt_dinov3", "models/fast_vlm", "models/flava", "models/fuyu", "models/idefics", "models/imagegpt", "models/kimi_k25", "models/layoutlmv2", "models/lfm2_vl", "models/llava", "models/minimax_m3_vl", "models/pix2struct"]
quantizations: []

[tarekziade]

Sorry I just realized that work was started earlier at #47218
I messed up when I did the triaging, sorry about that

[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.

[github-actions]

CI Results

Workflow Run ⚙️

Commit Info

Context	Commit	Description
RUN	61e10d08	workflow commit (merge commit)
PR	041d01ac	branch commit (from PR)
main	1f8daee0	base commit (on main)

✅ No failing test specific to this PR 🎉 👏 !