NeuroHD: Hunting Jackals, or Deploying and Accelerating Networks in VK Infrastructure
summary
NeuroHD is VK’s video-enhancement technology that increases resolution, raises frame rate, and reduces noise. It runs using two neural networks deployed and accelerated inside VK’s infrastructure to process video at scale.
AI-written summary · about
chapters (7)
- 0:00Introduction
- 3:26Project context and goals
- 7:42Rayf mesh: design and optimization
- 14:11Besik mesh: structure and challenges
- 20:39Infrastructure pitfalls
- 28:56Summary
- 29:38Q&A
related talks
33mArchitecting Scalable Multi-Modal Video Search
51mHuddly - The thinking camera and applied machine learning
22mAI Enhanced GPU Video Coding: Achieving Joint High Compression Efficiency and Throughput
20mScaling Edge AI with Elixir and Nerves
50mHow Neural Networks Helped Graphics
18mCoVA: Exploiting Compressed-Domain Analysis to Accelerate Video Analytics
06mMachines and the magic of fast learning
17mYuZu: Neural-Enhanced Volumetric Video Streaming