Problem with whisper-large-v3

Models
1

I was trying to use the new whisper-large-v3 model and get the following error:

File “/ext3/miniconda3/lib/python3.10/site-packages/transformers/tokenization_utils_base.py”, line 2065, in _from_pretrained
raise ValueError(
ValueError: Non-consecutive added token ‘<|0.02|>’ found. Should have index 50365 but has index 50366 in saved vocabulary.

Looks like some bug in the model?

2

I have met this question. Maybe need to upgrade the transformers package?

3

it didn’t help and i have the same problem

4

cc @ArthurZ maybe

5

Actually, it just fixed it for me…thanks!

6

upgrade the transformer package solve my problem

7

Hi did you fix it now? I meet the same problem :frowning:

8

Something changed in the whisper code. I just downgraded to 4.36 and now its working fine