Article URL: https://github.com/Video-Rebirth/hyperflow-h3-flash
Comments URL: https://news.ycombinator.com/item?id=50405263
Points: 1
# Comments: 0
AI & ML
Show HN: We compressed a 31B video+audio diffusion model to 14B, 6 steps, INT8
Hacker News New0 views
Source：https://github.com/Video-Rebirth/hyperflow-h3-flash
PreviousAn ode to livelyness analysis; SSA to registersNextWhy MySQL In-Memory Reads Are 50x Slower Than Redis (Even with 0% Disk I/O)
Related posts
The US let Nvidia sell chips to an empty office in China
0The Verge
NYT Cooking’s AI assistant saved my weeknight chicken, but almost ruined my cake
0The Verge
Costco members can get $5 off the reimagined Ocarina of Time
0The Verge
CVE-2026-101909 Impact Boundaries: Multipart, URL-Encoding, and Denial of Service
0dev.to
Google's "unusual traffic" CAPTCHA is about your IP, not you
0dev.to
Turning Solo AI Agents into an SRE Incident Squad with Switch
0dev.to