Whisper - get probability of detected language

### System Info - `transformers` version: 4.38.0.dev0 - Platform: Linux-4.15.0-142-generic-x86_64-with-glibc2.23 - Python version: 3.10.11 - Huggingface_hub version: 0.20.3 - Safetensors version: 0.4.2 - Accelerate version: 0.27.2 - Accelerate config: not found - PyTorch version (GPU?): 2.2.0+cu121 (True) - Tensorflow version (GPU?): 2.12.0 (True) ### Who can help? @sanchit-gandhi I guess, since he's the one who provided the answer in the previous git issue. ### 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 Following #25138, @sanchit-gandhi provided an answer to retrieve the language using Whisper model and processor (since Whisper conditionnal tokens include the language token). He later provided a little adaptation in order to get the probability of the language. This is a nice possibility. However, using the latest version of transformers it seems that it's not possible anymore (that's why I write it as a bug but could also be a feature request). Quick example in order to check : ```python language_identification = WhisperForConditionalGeneration.from_pretrained("openai/whisper-small").to("cuda:0") lid_processor = WhisperProcessor.from_pretrained("openai/whisper-small") audio, _ = librosa.load(<my_file>, sr=16000) lid = lid_processor(audio, sampling_rate=16000, return_tensors="pt", truncation=True) input_features = lid.input_features.to("cuda:0", torch.float32) outputs = language_identification.generate(input_features, output_scores=True, return_dict_in_generate=True, max_new_tokens=1) pred_text = lid_processor.batch_decode(outputs.sequences, skip_special_tokens=False) pred_text ``` `pred_text` is : ``` ['<|startoftranscript|><|en|><|transcribe|><|notimestamps|> 80'] ``` Here we see the conditionnal tokens as well as my only transcription token ` …