ZHZisZZ/dllm
dLLM: Simple Diffusion Language Modeling
Implements diffusion-based language models using masked and block diffusion approaches, with unified training pipelines built on HuggingFace `transformers` Trainer supporting LoRA, DeepSpeed, and FSDP. Provides evaluation harness integration and reference implementations for models like LLaDA, Dream, and BERT-Chat, plus tools for converting any autoregressive model to diffusion-based generation through adapter training.
2,193 stars.
Stars
2,193
Forks
206
Language
Python
License
Apache-2.0
Category
Last pushed
Feb 27, 2026
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/transformers/ZHZisZZ/dllm"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Compare
Related models
EnnengYang/Awesome-Model-Merging-Methods-Theories-Applications
Model Merging in LLMs, MLLMs, and Beyond: Methods, Theories, Applications and Opportunities. ACM...
pengzhangzhi/Open-dLLM
Open diffusion language model for code generation — releasing pretraining, evaluation,...
THUDM/LongWriter
[ICLR 2025] LongWriter: Unleashing 10,000+ Word Generation from Long Context LLMs
AIoT-MLSys-Lab/SVD-LLM
[ICLR 2025🔥] SVD-LLM & [NAACL 2025🔥] SVD-LLM V2
jxiw/MambaInLlama
[NeurIPS 2024] Official Repository of The Mamba in the Llama: Distilling and Accelerating Hybrid Models