GPU CLI

Run code on cloud GPUs with a single command.

# Local
python train.py

# Remote GPU - just add "gpu run"
gpu run python train.py

Use Cases

ML & Training

• Model training and fine-tuning
• LoRA/QLoRA for LLMs
• Batch inference
• Jupyter notebooks

Image Generation

• ComfyUI
• Automatic1111 / SD WebUI
• Fooocus
• InvokeAI
• Kohya SS (LoRA training)

Video & Audio

• Wan, AnimateDiff, CogVideo
• Whisper transcription
• Music generation

LLM Inference

• Ollama
• llama.cpp
• text-generation-webui

Why GPU CLI?

• No more forgotten pods - Auto-stops when idle, so you only pay for what you use
• Real-time outputs - Files sync back as they're created
• Just works - Your local files sync to the pod, outputs sync back

Quick Start

# Install
curl -fsSL https://gpu-cli.sh/install.sh | sh

# Run on remote GPU
gpu run python train.py

Examples

Run a training job:

gpu run python train.py --epochs 100

Spin up ComfyUI:

gpu run -p 8188:8188 python main.py --listen 0.0.0.0
# Opens http://localhost:8188 automatically

Interactive shell:

gpu run bash

Links

• Documentation
• Report Issues