Fix bad tensor shape in failing Hubert test.

[ebezzam]

What does this PR do?

Fixes breaking Hubert test: https://github.com/huggingface/transformers/actions/runs/16361063842/job/46229140250

cc @eustlb

Redoing the integrations test would be nice but original implementation is archived and installation/setup doesn't work because of configuration issues on their side that would take too much time... (see below traceback)

Otherwise fixed test is consistent with another one, see here

Traceback (most recent call last):
  File "scripts/test_hubert.py", line 32, in <module>
    model = SpeechEncoder.by_name(dense_model_name='mhubert-base-25hz', quantizer_model_name='kmeans',
  File "/home/eric_bezzam/transformers/textlesslib/textless/data/speech_encoder.py", line 134, in by_name
    dense_model = dispatch_dense_model(dense_model_name)
  File "/home/eric_bezzam/transformers/textlesslib/textless/__init__.py", line 38, in dispatch_dense_model
    return model_class(checkpoint_path, layer=model_layer, **kwargs)
  File "/home/eric_bezzam/transformers/textlesslib/textless/data/hubert_feature_reader.py", line 28, in __init__
    self.load_checkpoint_()
  File "/home/eric_bezzam/transformers/textlesslib/textless_env/lib/python3.8/site-packages/torch/utils/_contextlib.py", line 116, in decorate_context
    return func(*args, **kwargs)
  File "/home/eric_bezzam/transformers/textlesslib/textless/data/hubert_feature_reader.py", line 32, in load_checkpoint_
    model, _, task = fairseq.checkpoint_utils.load_model_ensemble_and_task(
  File "/home/eric_bezzam/transformers/textlesslib/textless_env/lib/python3.8/site-packages/fairseq/checkpoint_utils.py", line 421, in load_model_ensemble_and_task
    task = tasks.setup_task(cfg.task)
  File "/home/eric_bezzam/transformers/textlesslib/textless_env/lib/python3.8/site-packages/fairseq/tasks/__init__.py", line 39, in setup_task
    cfg = merge_with_parent(dc(), cfg)
  File "/home/eric_bezzam/transformers/textlesslib/textless_env/lib/python3.8/site-packages/fairseq/dataclass/utils.py", line 483, in merge_with_parent
    merged_cfg = OmegaConf.merge(dc, cfg)
  File "/home/eric_bezzam/transformers/textlesslib/textless_env/lib/python3.8/site-packages/omegaconf/omegaconf.py", line 321, in merge
    target.merge_with(*others[1:])
  File "/home/eric_bezzam/transformers/textlesslib/textless_env/lib/python3.8/site-packages/omegaconf/basecontainer.py", line 331, in merge_with
    self._format_and_raise(key=None, value=None, cause=e)
  File "/home/eric_bezzam/transformers/textlesslib/textless_env/lib/python3.8/site-packages/omegaconf/base.py", line 95, in _format_and_raise
    format_and_raise(
  File "/home/eric_bezzam/transformers/textlesslib/textless_env/lib/python3.8/site-packages/omegaconf/_utils.py", line 629, in format_and_raise
    _raise(ex, cause)
  File "/home/eric_bezzam/transformers/textlesslib/textless_env/lib/python3.8/site-packages/omegaconf/_utils.py", line 610, in _raise
    raise ex  # set end OC_CAUSE=1 for full backtrace
  File "/home/eric_bezzam/transformers/textlesslib/textless_env/lib/python3.8/site-packages/omegaconf/basecontainer.py", line 329, in merge_with
    self._merge_with(*others)
  File "/home/eric_bezzam/transformers/textlesslib/textless_env/lib/python3.8/site-packages/omegaconf/basecontainer.py", line 347, in _merge_with
    BaseContainer._map_merge(self, other)
  File "/home/eric_bezzam/transformers/textlesslib/textless_env/lib/python3.8/site-packages/omegaconf/basecontainer.py", line 305, in _map_merge
    dest._format_and_raise(key=key, value=src_value, cause=e)
  File "/home/eric_bezzam/transformers/textlesslib/textless_env/lib/python3.8/site-packages/omegaconf/base.py", line 95, in _format_and_raise
    format_and_raise(
  File "/home/eric_bezzam/transformers/textlesslib/textless_env/lib/python3.8/site-packages/omegaconf/_utils.py", line 694, in format_and_raise
    _raise(ex, cause)
  File "/home/eric_bezzam/transformers/textlesslib/textless_env/lib/python3.8/site-packages/omegaconf/_utils.py", line 610, in _raise
    raise ex  # set end OC_CAUSE=1 for full backtrace
omegaconf.errors.ValidationError: Value '50.0' could not be converted to Integer
	full_key: label_rate
	reference_type=Optional[HubertPretrainingConfig]
	object_type=HubertPretrainingConfig

[github-actions]

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

run-slow: hubert

[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.

[Rocketknight1]

Tested locally and it looks good, thank you for the PR!