Transformers 4.57.0 declares compatibility with Python>=3.9 but is only compatible with >=3.10

[anakin87]

System Info

• transformers version: 4.57.0
• Platform: macOS-15.7-arm64-arm-64bit
• Python version: 3.9.6
• Huggingface_hub version: 0.35.3
• Safetensors version: 0.6.2
• Accelerate version: 1.10.1
• Accelerate config: not found
• DeepSpeed version: not installed
• PyTorch version (accelerator?): 2.8.0 (NA)
• Tensorflow version (GPU?): not installed (NA)
• Flax version (CPU?/GPU?/TPU?): not installed (NA)
• Jax version: not installed
• JaxLib version: not installed
• Using distributed or parallel set-up in script?:

Who can help?

No response

Information

• The official example scripts
• My own modified scripts

Tasks

• An officially supported task in the examples folder (such as GLUE/SQuAD, ...)
• My own task or dataset (give details below)

Reproduction

1. Use Python 3.9 and do pip install -U "transformers[torch]"
2. from transformers import Trainer

Error:

Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "<frozen importlib._bootstrap>", line 1055, in _handle_fromlist
  File "/omitted/.venv/lib/python3.9/site-packages/transformers/utils/import_utils.py", line 2317, in __getattr__
    module = self._get_module(self._class_to_module[name])
  File "/omitted/.venv/lib/python3.9/site-packages/transformers/utils/import_utils.py", line 2347, in _get_module
    raise e
  File "/omitted/.venv/lib/python3.9/site-packages/transformers/utils/import_utils.py", line 2345, in _get_module
    return importlib.import_module("." + module_name, self.__name__)
  File "/Library/Developer/CommandLineTools/Library/Frameworks/Python3.framework/Versions/3.9/lib/python3.9/importlib/__init__.py", line 127, in import_module
    return _bootstrap._gcd_import(name[level:], package, level)
  File "/omitted/.venv/lib/python3.9/site-packages/transformers/trainer.py", line 311, in <module>
    class Trainer:
  File "/omitted/.venv/lib/python3.9/site-packages/transformers/trainer.py", line 5593, in Trainer
    def _get_num_items_in_batch(self, batch_samples: list, device: torch.device) -> int | None:
TypeError: unsupported operand type(s) for |: 'type' and 'NoneType'

Expected behavior

Currently, Transformers requires Python>=3.9 but only works with Python>=3.10

I see two possible solutions:

• Python 3.9 is reaching its End Of Life this month, so I think it can be dropped
• Modify the implementations to remove syntaxes that are only supported in >=3.10 (such as | instead of Union)

[SunMarc]

Sorry for that ! The issue should be fixed here #41359 and we will release a patch in the following days. Note that for the next release v5, we will drop support for py3.9

[anakin87]

@SunMarc feel free to close this if you think it's solved. Thanks!

[SunMarc]

I will close it when the patch release is out ! The main branch is not compatible with py3.9