CodeT5 tokenizer.model_max_length is 1000000000000000019884624838656

[khoda81]

System Info

- `transformers` version: 4.12.2
- Platform: Windows-10-10.0.22000-SP0
- Python version: 3.9.7
- PyTorch version (GPU?): 1.10.2 (True)
- Tensorflow version (GPU?): not installed (NA)
- Flax version (CPU?/GPU?/TPU?): not installed (NA)
- Jax version: not installed
- JaxLib version: not installed
- Using GPU in script?: False
- Using distributed or parallel set-up in script?: False

Who can help?

@patrickvonplaten

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

from transformers import AutoTokenizer

tokenizer = AutoTokenizer.from_pretrained("formermagic/codet5-large")
print(tokenizer.model_max_length)

This prints 1000000000000000019884624838656.

Expected behavior

To print 512.

[patrickvonplaten]

Hey @khoda81,

The bug seems to be related directly to the model of @mozharovsky and not really to a canonical model. However, also not that T5 does not have a maximum length. You can use T5 with lengths as long as you want until your memory errors out. T5 uses relative positional embeddings so that there is no static weight that will give you an "out-of-index" error.

Also see: #5204

[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.