Standardize BertGeneration model card

[nemitha2005]

What does this PR do?

#36979

Updated the BertGeneration model card to follow the new standardized format including:

• New consistent layout with badges
• Friendly description written for accessibility
• Usage examples with Pipeline, AutoModel, and transformers-cli
• Quantization example with BitsAndBytesConfig
• Updated resources section with proper links

This standardizes the BertGeneration documentation to match the new template format requested in issue #36979.

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](https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md#create-a-pull-request),
  Pull Request section?
• Was this discussed/approved via a Github issue or the [forum](https://discuss.huggingface.co/)? 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](https://github.com/huggingface/transformers/tree/main/docs), and
  [here are tips on formatting docstrings](https://github.com/huggingface/transformers/tree/main/docs#writing-source-documentation).
• Did you write any new necessary tests?

Who can review?

@stevhliu - Documentation lead who is managing the model card standardization project.

Anyone in the community is free to review the PR once the tests have passed.

[stevhliu]

Thanks!

[stevhliu]

Suggested change

	[BertGeneration](https://huggingface.co/papers/1907.12461) leverages pre-trained BERT checkpoints for sequence-to-sequence tasks using EncoderDecoderModel architecture.
	[BertGeneration](https://huggingface.co/papers/1907.12461) leverages pretrained BERT checkpoints for sequence-to-sequence tasks with the [`EncoderDecoderModel`] architecture. BertGeneration adapts the [`BERT`] for generative tasks.

[stevhliu]

Suggested change

BertGeneration adapts the powerful BERT encoder for generative tasks by using it in encoder-decoder architectures for tasks like summarization, translation, and text fusion. Think of it as taking BERT's deep understanding of language and teaching it to generate new text based on input context.

[stevhliu]

Suggested change

	You can find all the original BertGeneration checkpoints under the [BERT Generation](https://huggingface.co/models?search=bert-generation) collection.
	You can find all the original BERT checkpoints under the [BERT](https://huggingface.co/collections/google/bert-release-64ff5e7a4be99045d1896dbc) collection.

[stevhliu]

import torch
from transformers import pipeline

pipeline = pipeline(
    task="text2text-generation",
    model="google/roberta2roberta_L-24_discofuse",
    torch_dtype=torch.float16,
    device=0
)
pipeline("Plants create energy through ")

[stevhliu]

Lets show a pretrained example here:

import torch
from transformers import EncoderDecoderModel, AutoTokenizer

model = EncoderDecoderModel.from_pretrained("google/roberta2roberta_L-24_discofuse", torch_dtype="auto")
tokenizer = AutoTokenizer.from_pretrained("google/roberta2roberta_L-24_discofuse")

input_ids = tokenizer(
    "Plants create energy through ", add_special_tokens=False, return_tensors="pt"
).input_ids

outputs = model.generate(input_ids)
print(tokenizer.decode(outputs[0]))

[stevhliu]

Suggested change

	# Using transformers-cli for quick inference
	python -m transformers.models.bert_generation --model google/roberta2roberta_L-24_discofuse --input "This is the first sentence. This is the second sentence."
	echo -e "Plants create energy through " | transformers run --task text2text-generation --model "google/roberta2roberta_L-24_discofuse" --device 0

[stevhliu]

Suggested change

	The example below uses [BitsAndBytesConfig](../main_classes/quantization#transformers.BitsAndBytesConfig) to quantize the weights to 4-bit.
	The example below uses [BitsAndBytesConfig](../quantizationbitsandbytes) to quantize the weights to 4-bit.

[stevhliu]

Use the same code snippet with a pretrained model from the AutoModel example above

[stevhliu]

Suggested change

	- BertGenerationEncoder and BertGenerationDecoder should be used in combination with EncoderDecoderModel for sequence-to-sequence tasks.
	- [`BertGenerationEncoder`] and [`BertGenerationDecoder`] should be used in combination with [`EncoderDecoderModel`] for sequence-to-sequence tasks.
	<add code example here from https://huggingface.co/docs/transformers/model_doc/bert-generation#usage-examples-and-tips>

[stevhliu]

You can remove this section