Paper Detail

Dog-Breed Text-to-image Flow Matching from Pretraining to Fine-Tuning on 4 Consumer GPUs

Reyhaneh Esmailizadeh, Xingjian Leng, Zhanhao Liang, Liang Zheng

arxiv Score 5.8

Published 2026-10-13 · First seen 2026-10-14

General AI

Abstract

We generally perceive that training text-to-image diffusion models requires heavy compute: it is true if we aim at a comprehensive frontier model. This understanding prevents people from training their own models when getting started in the field. We develop MiniDog, a framework that enables pretraining diffusion transformers in 55 minutes and fine-tuning in another 15 minutes on four 3090 GPUs. For data, we use 26k dog images from ImageNet and caption them for pretraining and use 2k images generated by FLUX.1-schnell for fine-tuning. We precompute the image and text features and cache them for efficiency. For architecture, we use E2E-INVAE for its semantic latent space, a small transformer with 22M parameters, a small Qwen model for text encoding, and in-context conditioning. For training, we use iREPA and an EUPE perception encoder for acceleration. For evaluation, our repository supports FID, HPSv2 score, PickScore, and various VAE reconstruction metrics.

Workflow Status

Review status
pending
Role
unreviewed
Read priority
soon
Vote
Not set.
Saved
no
Collections
Not filed yet.
Next action
Not filled yet.

Reading Brief

No structured notes yet. Add `summary_sections`, `why_relevant`, `claim_impact`, or `next_action` in `papers.jsonl` to enrich this view.

Why It Surfaced

No ranking explanation is available yet.

Tags

No tags.

BibTeX

@article{esmailizadeh2026dog,
  title = {Dog-Breed Text-to-image Flow Matching from Pretraining to Fine-Tuning on 4 Consumer GPUs},
  author = {Reyhaneh Esmailizadeh and Xingjian Leng and Zhanhao Liang and Liang Zheng},
  year = {2026},
  abstract = {We generally perceive that training text-to-image diffusion models requires heavy compute: it is true if we aim at a comprehensive frontier model. This understanding prevents people from training their own models when getting started in the field. We develop MiniDog, a framework that enables pretraining diffusion transformers in 55 minutes and fine-tuning in another 15 minutes on four 3090 GPUs. For data, we use 26k dog images from ImageNet and caption them for pretraining and use 2k images gene},
  url = {https://arxiv.org/abs/2610.18870},
  keywords = {cs.CV},
  eprint = {2610.18870},
  archiveprefix = {arXiv},
}

Metadata

{}