Remove max_length from tokenization

[mreso]

What does this PR do?

Removes max_length from tokenization to prevent OOM
Fixes # (issue)

Feature/Issue validation/testing

Please describe the tests that you ran to verify your changes and relevant result summary. Provide instructions so it can be reproduced.
Please also list any relevant details for your test configuration.

• Test A
  Logs for Test A

echo "Hello" | CUDA_VISIBLE_DEVICES=0 python inference.py /fsx-project/shared/mreso/.cache/
huggingface/hub/models--meta-llama--Meta-Llama-3.1-8B-Instruct/snapshots/4281e96c7cf5ab6b312ef0cb78373efa3976a9dd/ --quantization '4bit' --enable_salesforce_content_safety False
use_fast_kernelsFalse
Loading checkpoint shards: 100%|████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 4/4 [00:22<00:00,  5.67s/it]
User prompt deemed safe.
User prompt:
Hello

Setting `pad_token_id` to `eos_token_id`:128001 for open-end generation.
the inference time is 6746.064721955918 ms
User input and model output deemed safe.
Model output:
Hello
Hello,
I'm still here, and I'm still loving you, in my own
fucking way.
--The Smiths
If we were supposed to be this happy and
contented, then surely we'd know it by now
--The Smiths
When I get to the end of my road
And the end of the road is my home
I find my heart still racing
With thoughts of you, as I roam
I'm going home
--Paul McCartney
If

Before submitting

• Did you read the contributor guideline,
  Pull Request section?
• Did you write any new necessary tests?

Thanks for contributing 🎉!

[albertodepaola]

LG!