Soup
Feinabstimmung von LLMs über eine einzige YAML-Datei. Layer-Streaming trainiert ein 8B-Modell auf einer 4-GB-Laptop-GPU.
Überblick
Anforderungen
Empfohlener VPS
Hostinger · KVM 4
4 vCPU · 16384 MB · 200 GB
Hostinger · KVM 4
4 vCPU · 16384 MB · 200 GB
Hostinger · KVM 8
8 vCPU · 32256 MB · 400 GB
Affiliate-Hinweis
Docker Compose
# Generated by Run This Ai — docker-compose.yml
services:
soup:
image: ghcr.io/makazhanalpamys/soup:latest
restart: unless-stopped
ports:
- 8080:8080
volumes:
- ./data/soup:/data
Verwandte Tools
NextChat
Cross-platform ChatGPT web UI with multi-model support for Ollama, Claude, Gemini, and more
Lobe Chat
Extensible, open-source ChatGPT alternative with plugins, knowledge base, and multi-LLM support
Open WebUI
User-friendly WebUI for LLMs (Ollama, OpenAI API)
text-generation-webui
Run local LLMs with a powerful web interface — text, vision, tool-calling, and OpenAI-compatible API
Streamlit
Build and share data apps in pure Python — fast
Gradio
Build and share delightful machine learning apps in Python
Anleitungen & Artikel
How to Fine-Tune an 8B Model on a 4 GB Laptop GPU with Soup
Step-by-step tutorial: install soup-cli, init a chat template, enable layer streaming, and train Llama-3.1-8B-Instruct on a 4 GB RTX 3050 Laptop with a 3.32 GB VRAM peak.
Soup: Fine-Tune LLMs from a Single YAML File
Soup is an open-source LLM fine-tuning toolkit that configures an entire training run from one YAML file. Its layer-streaming engine trains an 8B model on a 4 GB laptop GPU at 119.6 tok/s.