AttributeError: 'GPT2Model' object has no attribute '_get_initial_cache_position' in AudioLDM2Pipeline

[chaowenguo]

Describe the bug

! python3 -m pip install -U phonemizer
! sudo apt install -y --no-install-recommends espeak-ng
import scipy, transformers
import torch
from diffusers import AudioLDM2Pipeline

repo_id = "anhnct/audioldm2_gigaspeech"
pipe = AudioLDM2Pipeline.from_pretrained(repo_id, torch_dtype=torch.float16)
pipe = pipe.to("cuda")

# define the prompts
prompt = "An female actor say with angry voice"
transcript = "wish you have a good day, i hope you never forget me"
negative_prompt = "low quality"

# run the generation
audio = pipe(
    prompt,
    negative_prompt=negative_prompt,
    transcription=transcript,
    num_inference_steps=200,
    audio_length_in_s=8.0,
    num_waveforms_per_prompt=1,
    max_new_tokens=512
).audios

# save the best audio sample (index 0) as a .wav file
scipy.io.wavfile.write("techno_2.wav", rate=16000, data=audio[0])
from IPython.display import Audio
Audio("techno_2.wav")

keep showing 'GPT2Model' object has no attribute '_get_initial_cache_position's

the problem seems to happen in transformers, I downgrade to transformers==4.47.0 to make it works

Reproduction

see above

Logs

System Info

diffusers 0.34.0, transformers 4.53.3. I test it in kaggle p100

Who can help?

@yiyixuxu @DN6

[github-actions]

This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.

Please note that issues that do not follow the contributing guidelines are likely to be ignored.

[yiyixuxu]

is this still an issue with latest transformers?

[mafuba]

@yiyixuxu it seems to be still there