Ruurd Kuiper PRO

Ruurd

5 7 7

https://www.ruurdkuiper.com/

AI & ML interests

Diffusion for text, AI for healthcare

Recent Activity

updated a Space about 14 hours ago

Ruurd/semantic-harmonization-api

new activity about 15 hours ago

Ruurd/BYOD-tokenized-256:Add task_categories to metadata

new activity about 15 hours ago

Ruurd/LAD-training-1m-256:Add task category and repository links

View all activity

Organizations

updated a Space about 14 hours ago

Clinical Codelist API

🚀

New activity in Ruurd/BYOD-tokenized-256 about 15 hours ago

Add task_categories to metadata

#1 opened about 22 hours ago by

nielsr

New activity in Ruurd/LAD-training-1m-256 about 15 hours ago

Add task category and repository links

#2 opened about 22 hours ago by

nielsr

published a Space 1 day ago

Clinical Codelist API

🚀

posted an update 1 day ago

Post

🚀 Can you turn an existing LLM into a diffusion language model in just a few hours on a single GPU?

We introduce "BYOD: Build Your Own DLM", a lightweight approach to converting autoregressive LLMs into diffusion models using LoRA, without modifying the original model weights.

✅ Works with Llama, Gemma, Qwen, and Ministral (and more!)
✅ Just 3–4 GPU hours for 7–9B models
✅ Open-source training code and Google Colab notebook
✅ Interactive demos to try all four models!

We're particularly interested in making diffusion language model research more accessible and reproducible.

Paper, code, and demos:
BYOD: Build Your Own DLM - Efficient Autoregressive-to-Diffusion Language Model Conversion (2610.41574)

Would love to hear your feedback and ideas for improving these conversions!

@AyoubBagheri @Ruurd @UMCU