surogate
Train and fine-tune LLMs faster on any NVIDIA GPU
All Skills in This Repository (1)
Pure Emerald Level IndicatorsFrequently Asked Questions
FAQPage SchemaHow to install Surogate?▼
Run `npx skills add invergent-ai/surogate --all -g -y` in your terminal to install all tools in this suite globally.
What does Surogate do?▼
Surogate is a high-performance training engine for large language models, supporting full fine-tuning, LoRA, QLoRA, and reinforcement learning (GRPO and DPO) on NVIDIA GPUs.
Which GPUs does Surogate support?▼
It runs on most modern NVIDIA GPUs, from A100 and RTX 30-series up to H100, H200, and Blackwell B200 and RTX 50-series cards.
Can I fine-tune models with limited GPU memory?▼
Yes. Native CPU offloading and QLoRA options let you fine-tune large models at bf16 precision even with limited VRAM.
Do I need deep learning expertise to use Surogate?▼
No. You define training jobs in a simple YAML config file and launch them with a single command like `surogate sft config.yaml`.
Related Repositories in Software Engineering
View All in Software Engineering→
openclawopenclaw
Run a personal AI assistant across your devices and chat apps
obrasuperpowers
Gives coding agents a disciplined workflow from idea to merged code
facebookreact
AI agent skills for building, testing, and porting React core