Research map / Image generation / Text-to-image diffusion
Paragraph-to-Image Generation with Information-Enriched Diffusion Model
Weijia Wu, Zhuang Li, Yefei He, Mike Zheng Shou, Chunhua Shen, Lele Cheng, Yan Li, Tingting Gao, Di Zhang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add LoRA adaptation of the Llama V2 text encoder (ablation) | ParaDiffusion · ParaPrompts-400 | text faithfulness win rate: 53.4% win vs 44.5% lose | Improved |
| increase LoRA rank from 8 to 64 (ablation) | ParaDiffusion LoRA · ParaPrompts-400 | win rate: 50.6% -> 53.4% | Improved |
| increase LoRA rank from 64 to 128 (ablation) | ParaDiffusion LoRA · ParaPrompts-400 | win rate: 53.4% -> 51.9% | Hurt |
| LoRA alpha 32 instead of 16 (r=64) (ablation) | ParaDiffusion LoRA · ParaPrompts-400 | win rate: 52.3% -> 53.4% | Improved |
| LoRA alpha 64 instead of 32 (r=64) (ablation) | ParaDiffusion LoRA · ParaPrompts-400 | win rate: 53.4% -> 53.1% | No clear effect |
| lengthen captions from 10 to 150 words (ParaImage-Big 500k) (ablation) | ParaDiffusion · ParaPrompts-400 | text faithfulness voting: 8.5% -> 40.6% | Improved |
| lengthen captions from 10 to 150 words (ParaImage-Small 3k) (ablation) | ParaDiffusion · ParaPrompts-400 | text faithfulness voting: 16.4% -> 31.2% | Improved |
| train on manually curated ParaImage-Small instead of LAION (ablation) | ParaDiffusion · ParaPrompts-400 | visual appeal voting: 3.1% -> 54.7% | Improved |
| train on LAION-Aesthetics instead of LAION (ablation) | ParaDiffusion · ParaPrompts-400 | visual appeal voting: 3.1% -> 18.5% | Improved |
| increase ParaImage-Big size from 0.3m to 3.3m (ablation) | ParaDiffusion · MS-COCO 256x256 | FID-30K: 13.45 -> 9.64 | Improved |
| increase ParaImage-Big size from 0.3m to 3.3m (ablation) | ParaDiffusion · ParaPrompts-400 | visual appeal voting: 17.1% -> 50.2% | Improved |
| add Stage-2 training on ParaImage-Big (ablation) | ParaDiffusion · ParaPrompts-400 | visual appeal win: 82.1 win vs 17.9 lose | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 2 on most shared results.
- LLM Blueprint — LLM Blueprint: Enabling Text-to-Image Generation with Complex and Detailed Prompts (ICLR 2024) · better on 2 of 3 shared results
- RAPHAEL — RAPHAEL: Text-to-Image Generation via Large Mixture of Diffusion Paths (NeurIPS 2023) · better on 0 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Improving Long-Text Alignment for Text-to-Image Diffusion Models (ICLR 2025 · similarity 0.90)
- Noise Diffusion for Enhancing Semantic Faithfulness in Text-to-Image Synthesis (CVPR 2025 · similarity 0.90)
- Exploring the Deep Fusion of Large Language Models and Diffusion Transformers for Text-to-Image Synthesis (CVPR 2025 · similarity 0.90)
- LLM4GEN: Leveraging Semantic Representation of LLMs for Text-to-Image Generation (AAAI 2025 · similarity 0.89)
- Seg4Diff: Unveiling Open-Vocabulary Semantic Segmentation in Text-to-Image Diffusion Transformers (NeurIPS 2025 · similarity 0.89)
- LLM Blueprint: Enabling Text-to-Image Generation with Complex and Detailed Prompts (ICLR 2024 · similarity 0.89)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
