Dream 7B

### Model/Pipeline/Scheduler description

https://hkunlp.github.io/blog/2025/dream/

> In short, Dream 7B:
>
> - consistently outperforms existing diffusion language models by a large margin;
> - matches or exceeds top-tier Autoregressive (AR) language models of similar size on the general, math, and coding > abilities;
> - demonstrates strong planning ability and inference flexibility that naturally benefits from the diffusion modeling.

Basically a new SotA diffusion-based LLM. It would be great to introduce LLMs to the library's roster.

### Open source status

- [x] The model implementation is available.
- [x] The model weights are available (Only relevant if addition is not a scheduler).

### Provide useful links for the implementation

**Team**: Jiacheng Ye*, Zhihui Xie*, Lin Zheng*, Jiahui Gao*, Zirui Wu, Xin Jiang, Zhenguo Li, and Lingpeng Kong.
**Affiliations**: The University of Hong Kong, Huawei Noah’s Ark Lab

[Base HF Link](https://huggingface.co/Dream-org/Dream-v0-Base-7B)
[Instruction-tuned HF Link](https://huggingface.co/Dream-org/Dream-v0-Instruct-7B)
[Codebase](https://github.com/HKUNLP/Dream)

Start with the linked Dream codebase and the Base and Instruction-tuned Hugging Face model pages to understand the implementation and available weights. Done means the Dream 7B models are introduced to the diffusers library with their documented usage supported.

Scritto dal modello di indicizzazione a partire dal testo della issue.