Unsloth AI· @UnslothAI · X·· 2026-08-13AI 评分52
AI 导读
Unsloth 现已支持免费微调 Meta 的 30B 智能体模型 Muse Glimmer，并提供 Vision + Text、RL GRPO、Conversational 三类免费 Kaggle notebooks。其训练速度比 FA2 快约 1.5 倍、显存省约 50% 且无精度损失；30B QLoRA 可在 24GB VRAM 本地训练，LoRA 需 >40GB，也支持导出 GGUF 和全量微调（FFT）。指南：https://unsloth.ai/docs/models/muse-glimmer/train
正文 · 原文
You can now fine-tune Meta Muse Glimmer 30B for free! 🔥
Our free notebook also supports GRPO RL training.
Unsloth trains Muse Glimmer 1.5× faster with 50% less VRAM vs FA2 setups. Train locally with 24GB VRAM.
Guide: https://unsloth.ai/docs/models/muse-glimmer/train
Notebooks: https://unsloth.ai/docs/models/muse-glimmer/train#free-fine-tuning-notebooks
来源：Unsloth AI · x.com