README.md

metadata

datasets:
  - d3LLM/trajectory_data_dream_32
pipeline_tag: text-generation
library_name: transformers
license: apache-2.0
base_model: Dream-org/Dream-v0-Instruct-7B
tags:
  - diffusion
  - text-generation
  - fast-inference
  - d3llm

d3LLM: Ultra-Fast Diffusion LLM using Pseudo-Trajectory Distillation 🚀

This repository contains the d3LLM-Dream model, an ultra-fast diffusion language model introduced in the paper d3LLM: Ultra-Fast Diffusion LLM using Pseudo-Trajectory Distillation.

• 📄 Paper: arXiv:2601.07568
• 👉 Code repo: https://github.com/hao-ai-lab/d3LLM
• 🌐 Blog: https://hao-ai-lab.github.io/blogs/text-diffusion/
• 🕹️ Demo: https://d3llm-team.github.io/

Model Description

d3LLM-Dream is an ultra-fast diffusion language model that achieves high generation speed while maintaining competitive performance. It strikes a balance between accuracy and parallelism by using pseudo-trajectory distillation during training and entropy-based multi-block decoding during inference.

Key Features

• 🚀 High throughput: 4.5× faster than autoregressive models (Qwen-2.5-7B) on H100 GPU, 2.5× faster on A100 GPU. Achieves 235.34 tokens/s on H100 on GSM8K-CoT.
• 📊 High AUP: Optimized for Accuracy Under Parallelism across benchmarks.
• 🔧 Specialized: Optimized for coding and math reasoning tasks.

Usage

You can load and use the model with the 🤗 Transformers library. Note that trust_remote_code=True is required as the model uses a custom architecture.

from transformers import AutoModel, AutoTokenizer

model_id = "d3LLM/d3LLM_Dream"

# Load model and tokenizer
tokenizer = AutoTokenizer.from_pretrained(model_id, trust_remote_code=True)
model = AutoModel.from_pretrained(model_id, trust_remote_code=True)

# For detailed inference scripts (multi-block decoding), 
# please refer to the official GitHub repository.

For more comprehensive examples and evaluation scripts, visit the official repository.

Citation

@article{arxiv'26:d3llm,
  title   = {d3LLM: Ultra-Fast Diffusion LLM using Pseudo-Trajectory Distillation},
  author  = {Yu-Yang Qian and Junda Su and Lanxiang Hu and Peiyuan Zhang and Zhijie Deng and Peng Zhao and Hao Zhang},
  journal = {ArXiv preprint},
  volume  = {arXiv:2601.07568},
  year    = {2026}
}