Three diffusion papers head to ECCV: Flash-BoN, PFM and dynamic image-gen benchmarks
RisingSayak · x · 2026-09-07
RisingSayak is heading to ECCV in Sweden, starting with a tutorial on September 8, "Post-Training Diffusion Models: Enhancing Capabilities, Control, and Alignment," followed by three papers he co-authored:
- Flash-BoN: best-of-n is surprisingly effective when compute is spent better.
- PAFM: sampling posteriors better in flow-matching.
- DynEval: most image generation benchmarks are stale; making them dynamic lets models evolve without evolving the benchmark.
The works span post-training capability, control and evaluation for diffusion models, and the author invites attendees to challenge him on the details.
More from Multimodal
- Full animation short made with Google Gemini tools, 10,564 frames — AI_Andrew · 2026-10-26
- ostris pre-training "Primal VAE": 1-channel latents match modern 8x VAEs via patching — ostrisai · 2026-10-26
- Slopcore: AI-generated meme films turn low-fi slop aesthetics into a subculture — Yamapama · 2026-10-26
- TischEins, a large Midjourney SREF library, relaunches with new design after security outage — tisch_eins · 2026-10-26
- MOSS-Image-SFT 4-step distillation tops 16.5K downloads on Hugging Face — _akhaliq · 2026-10-26
- Claude Code can now edit videos: a 10-prompt workflow from raw footage to final cut — TawohAwa · 2026-10-26