Cartesia's H-Net ditches tokenization, learns dynamic byte grouping for up to 10x training efficiency
bycloud · youtube · 2026-11-23
A video explainer by bycloud on Cartesia's H-Net: traditional tokenization is suboptimal, and H-Net learns to dynamically group raw bytes instead, achieving up to 10x greater training efficiency on selected benchmarks while improving generalization to unfamiliar data. The approach may also extend to other modalities. Sources include Cartesia's blog series and the H-Net paper (arXiv:2507.07955).
More from Models
- Google's BuggyMLE injects bugs into ML repos, dropping top model solve rate from 95% to 9% — omarsar0 · 2026-11-23
- OpenAI chart reveals unreleased GPT-6 Astra models; Domingos questions billions for <0.3 bits — pmddomingos · 2026-11-23
- Replit agents speed up NumPy 18.9x — but only Claude's team scales — amasad · 2026-11-23
- Perplexity is a flawed cross-lab capability bound: tokenizer granularity favors OpenAI — davidmanheim · 2026-11-23
- Gemini 4 Hydrogen at $1.50/M Input Tokens Scores 51, Beating Argon's 50 — dl_weekly · 2026-11-23
- Base log perplexity is an extremely strong predictor of Epoch and OpenAI ECI, researchers note — stochasticchasm · 2026-11-23