Fine-tuned version of model - a raised exception

[egorsmkv]

Hello.

I am getting the following exception:

ValueError: Not fine-tuned model! Please, fine-tune the model first.

I have looked into the code and see that it needs to have Wav2Vec2ForPreTraining (self.model_config.architectures) in the ctc_finetuded_architectures variable.

Now this variable has these values:

{'WavLMForCTC', 'HubertForCTC', 'UniSpeechSatForCTC', 'Wav2Vec2ForCTC', 'UniSpeechForCTC', 'SEWForCTC', 'SEWDForCTC'}

I am running the code with this model - https://huggingface.co/Yehor/wav2vec2-xls-r-300m-uk-with-lm

I disabled the code that raises that exception and it seems there is no issue.

I would like to use some type of configuration to be able to run the code without changing the library code.

[jonatasgrosman]

Upgrade the library and try to reproduce the error again. I've just changed the code, and that issue is maybe fixed now.

[egorsmkv]

Now it works correctly. Thank you!