AINewsnow

Slang Is a Dial: Training a 125M Language Model From Scratch on Free GPUs

TL;DR : I built a language model from scratch, end to end, on free GPUs : a 125.9M-parameter model that writes in a register you choose, with slang, emoji, abbreviation, elongation and formality as five dials. The idea is simple and testable: measure the register of every training document with cod…

Read the full story at DEV Community — Machine Learning ↗

Timeline · 1 report

  1. 2026-11-02 11:47 · DEV Community — Machine Learning
    Slang Is a Dial: Training a 125M Language Model From Scratch on Free GPUs

More stories

  1. Malicious Claude Code extension pulled from Cursor marketplace after SSH backdoor attempt — Microsoft Research Blog
  2. DeepSeek open-sources dsh-superagent, an autonomous coding agent prototype — Bloomberg AI
  3. Anthropic updates Claude Code to block recreation of existing video games — Engadget
  4. Elon Musk says Grok Bot usage limits imposed as serving costs were 'unsustainably high' — Forbes AI
  5. Marvell to buy Lumentum in $115 billion all-stock AI data-center deal — CNBC Technology
  6. Anthropic files for IPO as Blackstone flags likely 'high' junk rating — Financial Times AI
  7. OpenAI pivots to 'AI compute campuses' and energy partnerships amid chip supply strain — Data Center Dynamics
  8. OpenAI lowers latest funding round target to $20 billion from $30 billion — Financial Times AI

Get the daily brief of stories like this at 6:30 every morning →