Miles v0.1.1 Ships Multi-LoRA Sharing and Native Claude Code Agentic Training
ying11231 · x · 2026-09-30
The open-source training framework Miles released v0.1.1 with several notable additions:
- Multi-LoRA with Tinker API compatibility: multiple training jobs can share one base model without interfering with each other.
- Agentic training: native integration with harnesses like Claude Code, and Harbor tasks can run directly in sandboxes such as AgentENV, Daytona, E2B and Modal.
- Lower memory for larger models: validated across NVIDIA and AMD GPUs.
- Stable-version model support: Qwen3.8-Flash-Next, GLM-5.3-Flash, Kimi-K3 and more.
More from coding & agent
- Developer Explores What AI Agents Should Actually Remember in Long-Term Memory Design — Bh0nu0077 · 2026-09-30
- OpenAI rolls out Codex cloud environments to Plus, Pro, Business and Enterprise — OpenAIDevs · 2026-09-30
- Why production agent builders say OpenAI's Agents API won't kill agent frameworks — BhAAI777 · 2026-09-30
- LlamaIndex's Jerry Liu and Snorkel AI on why evals and RL environments remain unsolved — ajratner · 2026-09-30
- Nous Research partners with OpenAI to bring Sign in with ChatGPT to Hermes Agent — NousResearch · 2026-09-30
- Remote is now Codex again, and Codex Cloud works like any other host — Dimillian · 2026-09-30