PyTorch Conf deck: how agents get trained in 2026 — SFT, distillation, GRPO, RL envs, TRL + OpenEnv
SergioPaniego · x · 2026-10-22
Slides from a talk at PyTorch Conf North America on how agents get trained in 2026 and how to do it yourself: SFT → distillation → GRPO → RL environments → training through the harness, all built with TRL + OpenEnv. The deck includes lots of links. Each of the first four steps also has a full episode with more detail in the author's summer Training Agents livestream series (playlist linked).
Related event: PyTorch Conf Slides Outline 2026 Agent Training Pipeline(2 posts)→
More from coding & agent
- Drop a PowerShell Script in a Folder, Get a Live MCP Tool — dfinke · 2026-10-23
- Auth'd MCP servers can deliver user-personalized skills, an overlooked design win — dsp_ · 2026-10-23
- Universe badge pairs RP2350 with an AGENTS.md so Copilot can code and deploy to it — mariorod1 · 2026-10-23
- Claude Code banked reset vanishes hours early, user demands Anthropic restore it — burny_tech · 2026-10-23
- Sudo L7 benchmark uses 76-word natural prompts vs 291-427 words for rivals — echen · 2026-10-23
- Claude recolors an AI-generated SVG into 4 sneaker colorways without touching stitching or shading — socialwithaayan · 2026-10-23