Research map / Probabilistic and scientific modeling / Variational and generative sampling
iVPF: Numerical Invertible Volume Preserving Flow for Efficient Lossless Compression
Check 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 |
|---|---|---|---|
| Use iVPF instead of IDF++ | Lossless flow compression · ImageNet32 | bpd: 4.12 -> 4.03 | Helped |
| Use iVPF instead of IDF++ | Lossless flow compression · ImageNet64 | bpd: 3.81 -> 3.75 | Helped |
| Use iVPF instead of IDF++ | Lossless flow compression · CIFAR10 | bpd: 3.26 -> 3.20 | Helped |
| Use iVPF instead of LBB | Lossless compression · CIFAR10 | bpd: 3.12 -> 3.20 | Hurt |
| Use iVPF instead of LBB | Lossless compression · ImageNet32 | bpd: 3.88 -> 4.03 | Hurt |
| Use iVPF instead of LBB | Lossless compression · ImageNet64 | bpd: 3.70 -> 3.75 | Hurt |
| Replace LBB with iVPF | CIFAR10 lossless compression · CIFAR10 | coding time: 97 -> 6.6 ms | Helped |
| Replace LBB with iVPF | CIFAR10 lossless compression · CIFAR10 | auxiliary bits: 39.86 -> 6.00 | Helped |
| Use iVPF instead of FLIF | High-resolution image compression · CLIC.mobile | bpd: 2.49 -> 2.47 | Helped |
| Use iVPF instead of FLIF | High-resolution image compression · CLIC.pro | bpd: 2.78 -> 2.63 | Helped |
| Use iVPF instead of PNG | High-resolution image compression · DIV2k | bpd: 3.09 -> 2.77 | Helped |
| Evaluate the ImageNet64-trained iVPF model rather than the ImageNet32-trained model | iVPF · CLIC.mobile | bpd: 2.47 -> 2.39 | Helped |
Closest papers by meaning
- In defense of parameter sharing for model-compression (ICLR 2024 · similarity 0.84)
- Lossy Compression with Pretrained Diffusion Models (ICLR 2025 · similarity 0.83)
- SlimSAM: 0.1% Data Makes Segment Anything Slim (NeurIPS 2024 · similarity 0.83)
- Compression via Pre-trained Transformers: A Study on Byte-Level Multimodal Data (ICML 2025 · similarity 0.83)
- 70% Size, 100% Accuracy: Lossless LLM Compression for Efficient GPU Inference via Dynamic-Length Float (DFloat11) (NeurIPS 2025 · similarity 0.83)
- Accelerating Parallel Diffusion Model Serving with Residual Compression (NeurIPS 2025 · similarity 0.83)
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.