Update modernbertdecoder docs

[orionw]

What does this PR do?

Updates the docs for the ModernBERTDecoder class to link to the paper and to the correct model name.

Before submitting

• This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case).
• Did you read the contributor guideline,
  Pull Request section?
• Was this discussed/approved via a Github issue or the forum? Please add a link
  to it if that's the case.
• Did you make sure to update the documentation with your changes? Here are the
  documentation guidelines, and
  here are tips on formatting docstrings.
• Did you write any new necessary tests?

Who can review?

Anyone in the community is free to review the PR once the tests have passed. Feel free to tag
members/contributors who may be interested in your PR.

@stevhliu for docs, if you have a sec!

[stevhliu]

Thanks for updating, I think you can remove the ## Usage tips section at the bottom since most of the information is already incorporated in the card.

It may also be nice to include a quantization example with the ettin-decoder-1b checkpoint!

[orionw]

Thanks so much for the suggestions @stevhliu, added them!

It may also be nice to include a quantization example with the ettin-decoder-1b checkpoint!

I actually am not super good at quantization (others were super kind and contributed an onyx version to some models that IDK how to do), so I might just leave it for now - although I love the idea!

[stevhliu]

The snippet below should work for quantization! :)

import torch
from transformers import AutoModelForCausalLM, AutoTokenizer, BitsAndBytesConfig

quantization_config = BitsAndBytesConfig(
    load_in_8bit=True,
)

tokenizer = AutoTokenizer.from_pretrained("jhu-clsp/ettin-decoder-1b")
model = AutoModelForCausalLM.from_pretrained(
    "jhu-clsp/ettin-decoder-1b",
    torch_dtype=torch.float16,
    device_map="auto",
    quantization_config=quantization_config
)

prompt = "The future of artificial intelligence is"
inputs = tokenizer(prompt, return_tensors="pt").to("cuda")

with torch.no_grad():
    outputs = model.generate(
        **inputs,
        max_length=50,
        num_return_sequences=1,
        temperature=0.7,
        do_sample=True,
        pad_token_id=tokenizer.eos_token_id
    )

generated_text = tokenizer.decode(outputs[0], skip_special_tokens=True)
print(f"Generated text: {generated_text}")

[orionw]

Awesome, thank you so much @stevhliu, added that to the model card!

Is there anything else to add for this PR or is it good to go?

[HuggingFaceDocBuilderDev]

The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.

[stevhliu]

I think for this PR, we should still:

1. remove the ## Usage tips section at the bottom since most of the information is already incorporated in the card
2. add the quantization example to this doc as well

[orionw]

Perfect, sorry I missed that. Added now!

[stevhliu]

ty ❤️