Hugging Face hosts quantization livestream with vLLM, bartowski, and LiquidAI maintainers
unofficialmerve · reddit · 2026-11-05
- Hugging Face's Merve Noyan is hosting a Local AI 301 quantization workshop livestream (Nov 10, 9AM PST / 6PM CET) on YouTube, X, and LinkedIn.
- Four talks + live Q&A: Merve on quantization basics (double quantization, calibration), bartowski (Colin Kealty, RedHat) on calibration-based quantization, Maxime Labonne & Aditya Tadimeti (LiquidAI) on quantization-aware distillation, and vLLM lead maintainer Michael Goin (RedHat) on fast microscaling formats on any GPU.
- Covers llama.cpp k-quants/i-quants; beginner-friendly.
Related event: Hugging Face to Host Local AI 301 Livestream on Model Quantization(4 posts)→
More from Infra
- Trainium's biggest hurdle isn't the chip — it's the CUDA ecosystem — DavidLinthicum · 2026-11-05
- Nebius auto-tuner speeds RL training up to 5x on fixed GPU budgets — demian_ai · 2026-11-05
- After communities vote against data centers, OpenAI pledges to pay power bills — ross2000 · 2026-11-05
- DeepSeek V4.1 streaming EXL3 packs bring ~70 tok/s local inference to 128GB Macs — HankYeomans · 2026-11-05
- 137 internal Google schemas reveal exactly what NavBoost remembers about every search — jedisct1 · 2026-11-05
- After 27 years of self-hosting, blogger shutters his own servers — genmon · 2026-11-05