Research map / Image generation / Fast diffusion sampling
TR0N: Translator Networks for 0-Shot Plug-and-Play Conditional Generation
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 TR0N class conditioning with a stochastic translator and Langevin error correction | NVAE · CIFAR-10 | FID: 41.70 -> 19.79 | Helped |
| Add TR0N class conditioning with a stochastic translator and Langevin error correction | AutoGAN · CIFAR-10 | FID: 12.45 -> 10.69 | Helped |
| Add TR0N with StyleGAN-XL, LAION2BCLIP, and BLIP-2 | Unconditional StyleGAN-XL · MS-COCO | FID: 10.9 FID | Helped |
| Add Langevin error correction (ablation) | TR0N:NVAE+ResNet50 · CIFAR-10 | FID: 40.80 -> 19.79 | Helped |
| Add the translator (ablation) | TR0N:NVAE+ResNet50 · CIFAR-10 | FID: 36.74 -> 19.79 | Helped |
| Use a stochastic rather than deterministic translator (ablation) | TR0N:NVAE+ResNet50 · CIFAR-10 | FID: 77.11 -> 19.79 | Helped |
| Use gradient descent rather than ADAM for latent updates (ablation) | TR0N:NVAE+ResNet50 · CIFAR-10 | FID: 20.24 -> 19.79 | Helped |
| Add Langevin error correction (ablation) | TR0N:AutoGAN+ResNet50 · CIFAR-10 | FID: 11.00 -> 10.69 | Helped |
| Add the translator (ablation) | TR0N:AutoGAN+ResNet50 · CIFAR-10 | FID: 14.30 -> 10.69 | Helped |
| Use a stochastic rather than deterministic translator (ablation) | TR0N:AutoGAN+ResNet50 · CIFAR-10 | FID: 123.23 -> 10.69 | Helped |
| Use gradient descent rather than ADAM for latent updates (ablation) | TR0N:AutoGAN+ResNet50 · CIFAR-10 | FID: 11.08 -> 10.69 | Helped |
| Use the fully Bayesian energy weights beta1=1 and beta2=1 rather than the classifier-focused setting (ablation) | TR0N:NVAE+ResNet50 · CIFAR-10 | FID: 19.79 -> 20.12 | Hurt |
Closest papers by meaning
- Structured Generative Adversarial Networks (NeurIPS 2017 · similarity 0.91)
- Plug & Play Generative Networks: Conditional Iterative Generation of Images in Latent Space (CVPR 2017 · similarity 0.89)
- Taming the Tail in Class-Conditional GANs: Knowledge Sharing via Unconditional Training at Lower Resolutions (CVPR 2024 · similarity 0.88)
- Attack Deterministic Conditional Image Generative Models for Diverse and Controllable Generation (AAAI 2024 · similarity 0.87)
- Turning Waste into Wealth: Leveraging Low-Quality Samples for Enhancing Continuous Conditional Generative Adversarial Networks (AAAI 2024 · similarity 0.87)
- Discovering Latent Graphs with GFlowNets for Diverse Conditional Image Generation (NeurIPS 2025 · similarity 0.87)
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.