register_quantizer or register_quantization_config does not add new method to QuantizationMethod

[Enlion91]

System Info

transformers==4.51.3
Ascend 910B
torch==2.5.1
torch-npu==2.5.1

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

I want to use transformers/quantizers/auto.py::register_quantization_config and register_quantizer functions to add my own quantizaiton method, but both of them would not add new method enum to the transformers/utils/quantization_config.py::QuantizationMethod, which is used to check if the new method is valid.
Therefore, thre register functions could not register the new method corretly. Users have to write code to manually add Enum to the QuantizationMethod.

Expected behavior

After registers functions called, the new method item should be added to QuantizationMethod as well.

@SunMarc @MekkCyber @ivarflakstad

[SunMarc]

Hey @Enlion91 , mind sharing the traceback of the error you are getting ? Also, not sure if you saw but there is an example of this feature here:

transformers/examples/quantization/custom_quantization_int8_example.py

Line 3 in 051a8ac

[Enlion91]

hi, sorry for this false bug report, which I encountered during my debug process.
QuantizationMethod with str works.
thank you for your reply!