Picture for Songhua Liu

Songhua Liu

Ultra-Resolution Adaptation with Ease

Add code
Mar 20, 2025
Viaarxiv icon

POSTA: A Go-to Framework for Customized Artistic Poster Generation

Add code
Mar 19, 2025
Viaarxiv icon

OminiControl2: Efficient Conditioning for Diffusion Transformers

Add code
Mar 11, 2025
Viaarxiv icon

Understanding Dataset Distillation via Spectral Filtering

Add code
Mar 03, 2025
Viaarxiv icon

Rethinking Large-scale Dataset Compression: Shifting Focus From Labels to Images

Add code
Feb 10, 2025
Figure 1 for Rethinking Large-scale Dataset Compression: Shifting Focus From Labels to Images
Figure 2 for Rethinking Large-scale Dataset Compression: Shifting Focus From Labels to Images
Figure 3 for Rethinking Large-scale Dataset Compression: Shifting Focus From Labels to Images
Figure 4 for Rethinking Large-scale Dataset Compression: Shifting Focus From Labels to Images
Viaarxiv icon

CLEAR: Conv-Like Linearization Revs Pre-Trained Diffusion Transformers Up

Add code
Dec 20, 2024
Viaarxiv icon

One-shot Federated Learning via Synthetic Distiller-Distillate Communication

Add code
Dec 06, 2024
Viaarxiv icon

OminiControl: Minimal and Universal Control for Diffusion Transformer

Add code
Nov 25, 2024
Figure 1 for OminiControl: Minimal and Universal Control for Diffusion Transformer
Figure 2 for OminiControl: Minimal and Universal Control for Diffusion Transformer
Figure 3 for OminiControl: Minimal and Universal Control for Diffusion Transformer
Figure 4 for OminiControl: Minimal and Universal Control for Diffusion Transformer
Viaarxiv icon

Teddy: Efficient Large-Scale Dataset Distillation via Taylor-Approximated Matching

Add code
Oct 10, 2024
Viaarxiv icon

LinFusion: 1 GPU, 1 Minute, 16K Image

Add code
Sep 03, 2024
Viaarxiv icon