XiangLi1999/Diffusion-LM

Health v2 · maintenance only

Flags: no_releases

Adoption not part of the score

1245 stars · 164 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

Diffusion-LM is the official research code for the paper 'Diffusion-LM Improves Controllable Text Generation', implementing a diffusion-based language model over continuous embeddings. It provides training, decoding, and classifier-guided generation pipelines for controllable text generation tasks.

Use cases

• generate text with diffusion models
• control text generation with syntactic constraints
• train a diffusion language model on custom text data
• reproduce results from the Diffusion-LM paper
• guide text generation with a classifier
• infill text conditioned on attributes like syntax trees

When to choose

• you need controllable text generation via diffusion
• you want to experiment with or extend diffusion-based language models
• you are reproducing or building on the Diffusion-LM research

When to avoid

• you need a production-ready text generation system
• you want simple autoregressive LLM inference without research setup
• you lack GPU resources for training diffusion models

Facets

library · maturity maintenance

machine-learning deep-learning nlp llm-training machine-learning deep-learning python diffusion-models controllable-text-generation research-code text-generation academic-paper natural-language-processing research linux gpu

Member repositories

For agents

markdown · JSON · MCP: product_card(name="XiangLi1999/Diffusion-LM")