Together open-sources Post Training Cookbook with GRPO, SFT and Multi-LoRA recipes
togethercompute · x · 2026-10-22
Together launched an open-source post-training-cookbook showcasing its Post Training service, which keeps the training loop in your code—you define data, rewards, loss, optimizer, and LoRA or full-weight updates—while GPU provisioning and infrastructure are handled by the platform.
The Apache-2.0 repo includes starter notebooks such as GRPO training on GSM8K, supervised fine-tuning, and a Multi-LoRA example running two adapters. Setup uses uv sync --locked with TOGETHERAPIKEY, HFTOKEN for gated models, and optional WANDBAPIKEY.
More from coding & agent
- OpenClaw opens its team environment to community contributors via Focus Weeks — RichardSocher · 2026-10-22
- ESA's cat keyboard incident is a warning for leaving AI agents unattended — vincent_koc · 2026-10-22
- Defense must run at machine speed: agents patch, engineers judge — saranormous · 2026-10-22
- TraceForge synthesizes reasoning traces from answers, beating Claude's own CoT by +3.7/+7.1 — IndependentResearcher121 · 2026-10-22
- Internet Archive fends off PHP attacks overnight, using Claude to patch old code — kingslyj · 2026-10-22
- Combining Ghidra, AI Reverse Engineering and WebGPU Sparks an "Absolute Renaissance" — thursdai_pod · 2026-10-22