mlsanityData to 13 Oct 2026

FlowLM: Few-Step Language Modeling via Diffusion-to-Flow Adaptation

arXiv 2605.20199 ↗First posted 6 Apr 2026cs.CL cs.AIcited by 0

On the map: Diffusion models → Discrete and text diffusion

Abstract (arXiv)

We present FlowLM, a flow matching language model transformed from pre-trained diffusion language models via efficient fine-tuning. By re-aligning the curved sampling trajectories of diffusion models into straight-line flows, FlowLM enables high quality few-step generation that rivals or even outperforms the quality of 2,000-step diffusion sampling with very few training epochs. Remarkably, finetuned FlowLM reaches performance saturation with only half as many training epochs as training from scratch, both approaches greatly outperforming the original diffusion model, thereby validating our method. Furthermore, we validate a more effective training objective for flow matching: predicting clean data to consistently guide the sampling process towards the true data distribution. Empirical results demonstrate that our approach is highly effective for high-quality, few-step text generation.

Key facts. No citing papers in the map yet (data through 13 Oct 2026). Placed in Discrete and text diffusion, one of 14 directions in Diffusion models. Builds on 26 of its 32 references that are in the map. First posted on arXiv 6 Apr 2026.

Builds on

26 of 32 references.

Closest ancestors · 9

Same direction, newest first

  1. 2025Dream-Coder 7B: An Open Diffusion Language Model for Code1 Sep
  2. 2025DLM-One: Diffusion Language Models for One-Step Sequence Generation30 May
  3. 2025TESS 2: A Large-Scale Generalist Diffusion Language Model19 Feb
  4. 2025Large Language Diffusion Models14 Feb
  5. 2024Scaling Diffusion Language Models via Adaptation from Autoregressive Models23 Oct
  6. 2024Simple and Effective Masked Diffusion Language Models11 Jun
  7. 2023Discrete Diffusion Modeling by Estimating the Ratios of the Data Distribution25 Oct
  8. 2023DiffuSeq-v2: Bridging Discrete and Continuous Text Spaces for Accelerated Seq2Seq Diffusion Models9 Oct
  9. 2022DiffuSeq: Sequence to Sequence Text Generation with Diffusion Models17 Oct

Other references · 17

  1. 2020Denoising Diffusion Probabilistic Modelscited by 20,952
  2. 2023GPT-4 Technical Reportcited by 18,278
  3. 2024The Llama 3 Herd of Modelscited by 14,486
  4. 2022Flow Matching for Generative Modelingcited by 5,542
  5. 2019BERTScore: Evaluating Text Generation with BERTcited by 5,480
  6. 2022Flow Straight and Fast: Learning to Generate and Transfer Data with Rectified Flowcited by 3,607
  7. 2022DPM-Solver: A Fast ODE Solver for Diffusion Probabilistic Model Sampling in Around 10 Stepscited by 1,974
  8. 2022Diffusion-LM Improves Controllable Text Generationcited by 1,012
  9. 2025Fast-dLLM: Training-free Acceleration of Diffusion LLM by Enabling KV Cache and Parallel Decodingcited by 409
  10. 2025LLaDA 1.5: Variance-Reduced Preference Optimization for Large Language Diffusion Modelscited by 275
  11. 2024Score identity Distillation: Exponentially Fast Distillation of Pretrained Diffusion Models for One-Step Generationcited by 212
  12. 2020Neural CRF Model for Sentence Alignment in Text Simplificationcited by 106
5 more
  1. 2025Improved Mean Flows: On the Challenges of Fastforward Generative Modelscited by 106
  2. 2024PeRFlow: Piecewise Rectified Flow as Universal Plug-and-Play Acceleratorcited by 97
  3. 2017Quasar: Datasets for Question Answering by Search and Readingcited by 50
  4. 2024Adversarial Score identity Distillation: Rapidly Surpassing the Teacher in One Stepcited by 47
  5. 2025Online Reward-Weighted Fine-Tuning of Flow Matching with Wasserstein Regularizationcited by 42

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2025FS-DFM: Fast and Accurate Long Text Generation with Few-Step Diffusion Language Models
  2. 2026Flow Map Language Models: One-step Language Modeling via Continuous Denoising
  3. 2026LangFlow: Continuous Diffusion Rivals Discrete in Language Modeling
  4. 2026dLLM: Simple Diffusion Language Modeling
  5. 2026Towards Faster Language Model Inference Using Mixture-of-Experts Flow Matching
  6. 2026Diffusion Language Models: An Experimental Analysis
  7. 2024Language Rectified Flow: Advancing Diffusion Language Generation with Probabilistic Flows
  8. 2026Simple Diffusion Language Models Are More Effective Few-Step Generators Than Reported