Getting error RecursionError: maximum recursion depth exceeded anytime trying to download a SequenceToSequence model

[Daremitsu1]

Describe the current behavior
While trying to download a SequenceToSequence Model from hugging face, be it T5 or Bart, getting the error RecursionError: maximum recursion depth exceeded.

Describe the expected behavior
model = SequenceToSequence(data,backbone='t5-small') when this is run then the model should get downloaded and can be used for further refinement.

`resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
---------------------------------------------------------------------------
RecursionError                            Traceback (most recent call last)
/tmp/ipykernel_12566/1708539698.py in <cell line: 1>()
----> 1 model = SequenceToSequence(data,backbone='t5-small')

24 frames
... last 1 frames repeated, from the frame below ...

/usr/lib/python3.10/ssl.py in verify_mode(self, value)
    736     @verify_mode.setter
    737     def verify_mode(self, value):
--> 738         super(SSLContext, SSLContext).verify_mode.__set__(self, value)
    739 
    740 

RecursionError: maximum recursion depth exceeded

Instead getting above error.

What web browser you are using
Chrome

Additional context
Link to a minimal, public, self-contained notebook that reproduces this issue.

• Share the file using your GitHub account using File > Save a copy as a GitHub Gist.
• or Share Drive notebooks using the Share button then 'Get Shareable Link'.

[EvanWiederspan]

We'll need a minimal reproducible notebook in order to diagnose more, but if you are running someone else's notebook, you will likely have better luck reaching out to the notebook author instead

[Daremitsu1]

@EvanWiederspan providing notebook link
https://github.com/Daremitsu1/Reproducible-code/blob/master/question_generation_from_compliance_with_sequencetosequence.ipynb

[katlyn-edwards]

I don't know if we can use that notebook -- it relies on specific data files uploaded that we don't have, such as:

input_csv = '/content/FHA Single Family Servicing Guide – 4000.1- Published 01-18-2023 - FINAL UPLOAD.csv'
train_csv = '/content/train.csv'
test_csv = '/content/test.csv'

I think you'll have more luck posting on StackOverflow, or reaching out to the maintainer of SequenceToSequence

[Daremitsu1]

@katlyn-edwards will a github link to the aforementioned file do?
https://github.com/Daremitsu1/Reproducible-code
Added both code as well as the file for reference