Instructions to use Chang-chih/leviathan-16B-final with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- llama-cpp-python
How to use Chang-chih/leviathan-16B-final with llama-cpp-python:
# !pip install llama-cpp-python from llama_cpp import Llama llm = Llama.from_pretrained( repo_id="Chang-chih/leviathan-16B-final", filename="leviathan-16B-final.gguf", )
llm.create_chat_completion( messages = "No input example has been defined for this model task." )
- Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- llama.cpp
How to use Chang-chih/leviathan-16B-final with llama.cpp:
Install (macOS, Linux)
curl -LsSf https://llama.app/install.sh | sh # Start a local OpenAI-compatible server with a web UI: llama serve -hf Chang-chih/leviathan-16B-final # Run inference directly in the terminal: llama cli -hf Chang-chih/leviathan-16B-final
Install from WinGet (Windows)
winget install llama.cpp # Start a local OpenAI-compatible server with a web UI: llama serve -hf Chang-chih/leviathan-16B-final # Run inference directly in the terminal: llama cli -hf Chang-chih/leviathan-16B-final
Use pre-built binary
# Download pre-built binary from: # https://github.com/ggerganov/llama.cpp/releases # Start a local OpenAI-compatible server with a web UI: ./llama-server -hf Chang-chih/leviathan-16B-final # Run inference directly in the terminal: ./llama-cli -hf Chang-chih/leviathan-16B-final
Build from source code
git clone https://github.com/ggerganov/llama.cpp.git cd llama.cpp cmake -B build cmake --build build -j --target llama-server llama-cli # Start a local OpenAI-compatible server with a web UI: ./build/bin/llama-server -hf Chang-chih/leviathan-16B-final # Run inference directly in the terminal: ./build/bin/llama-cli -hf Chang-chih/leviathan-16B-final
Use Docker
docker model run hf.co/Chang-chih/leviathan-16B-final
- LM Studio
- Jan
- Ollama
How to use Chang-chih/leviathan-16B-final with Ollama:
ollama run hf.co/Chang-chih/leviathan-16B-final
- Unsloth Studio
How to use Chang-chih/leviathan-16B-final with Unsloth Studio:
Install Unsloth Studio (macOS, Linux, WSL)
curl -fsSL https://unsloth.ai/install.sh | sh # Run unsloth studio unsloth studio -H 0.0.0.0 -p 8888 # Then open http://localhost:8888 in your browser # Search for Chang-chih/leviathan-16B-final to start chatting
Install Unsloth Studio (Windows)
irm https://unsloth.ai/install.ps1 | iex # Run unsloth studio unsloth studio -H 0.0.0.0 -p 8888 # Then open http://localhost:8888 in your browser # Search for Chang-chih/leviathan-16B-final to start chatting
Using HuggingFace Spaces for Unsloth
# No setup required # Open https://huggingface.co/spaces/unsloth/studio in your browser # Search for Chang-chih/leviathan-16B-final to start chatting
- Atomic Chat new
- Docker Model Runner
How to use Chang-chih/leviathan-16B-final with Docker Model Runner:
docker model run hf.co/Chang-chih/leviathan-16B-final
- Lemonade
How to use Chang-chih/leviathan-16B-final with Lemonade:
Pull the model
# Download Lemonade from https://lemonade-server.ai/ lemonade pull Chang-chih/leviathan-16B-final
Run and chat with the model
lemonade run user.leviathan-16B-final-{{QUANT_TAG}}List all available models
lemonade list
🐋 Leviathan-16B-Final: 旗艦參謀 · 師祖
模型簡介 (Model Introduction)
leviathan-16B-final 是利維坦艦隊的旗艦參謀(師祖),是艦隊中最高層級的戰略分析核心。本模型以 deepseek-ai/DeepSeek-V2-Lite-Chat 為基底,經過嚴格的蒸餾與量化流程,具備卓越的戰略推理、兵法應用與數位轉型分析能力。它是艦隊中所有後繼模型的知識源頭與行為標竿。
血統與煉製 (Lineage & Refinement)
| 項目 | 說明 |
|---|---|
| 基底模型 | deepseek-ai/DeepSeek-V2-Lite-Chat (MoE 架構) |
| 蒸餾來源 | 基於 DeepSeek-V2-Lite-Chat 的知識蒸餾 |
| 量化格式 | Q4_K_M (GGUF) |
| 檔案大小 | 10 GB |
| 部署狀態 | 已部署於 Ollama,相容於 llama.cpp |
艦隊定位 (Fleet Position)
| 面向 | 說明 |
|---|---|
| 角色 | 師祖(旗艦參謀) |
| 核心能力 | 孫子兵法戰略分析、數位轉型決策、高複雜度推理 |
| 回應結構 | 強制遵循「分析→判斷→建議」之參謀框架 |
| 繼承者 | Chang-chih/leviathan-10.5b-final (主力旗艦) 繼承其行為邏輯與知識 |
硬體相容性與需求 (Hardware Compatibility)
為確保模型能順利運行,請確認你的環境符合以下需求:
| 組件 | 最低需求 | 建議配置 |
|---|---|---|
| 顯存 (VRAM) | 16 GB (4-bit 量化) | 24 GB 以上 |
| 系統記憶體 (RAM) | 32 GB | 64 GB 以上 |
| 儲存空間 | 15 GB | 20 GB (含下載與快取) |
| GPU | 支援 CUDA 的 NVIDIA GPU | RTX 40 系列或更高 |
| 軟體環境 | Python 3.10+, Ollama / llama.cpp | 最新版 Ollama |
免責聲明 (Disclaimer)
- 使用責任:本模型僅供研究與技術分享使用。使用者應對其生成的內容自行負責,並遵守當地法律法規。
- 輸出準確性:模型可能產生不準確、偏誤或不適當的內容。不應將其作為專業決策(如醫療、法律、金融)的唯一依據。
- 開源社群:作為開源專案,我們歡迎社群反饋與改進,但無法保證模型在所有環境下的絕對穩定性。
致謝 (Acknowledgement)
本模型的開發與公開,得益於開源社群的卓越貢獻與支持:
- DeepSeek:作為此模型的邏輯與架構骨幹,提供了強大的基礎模型。
- QLoRA 與 bitsandbytes 團隊:提供高效微調與量化技術。
- Hugging Face 社群:提供模型分享與協作平台。
- 利維坦計畫的開發者與測試者:感謝所有參與者,特別是
Chang-chih的貢獻與付出。
部署方式 (Deployment)
透過 Ollama (推薦)
# 直接運行
ollama run Chang-chih/leviathan-16B-final "你的問題"
# 或先下載模型
ollama pull Chang-chih/leviathan-16B-final
透過 llama.cpp
bash
# 下載 GGUF 檔案後
./llama-cli -m leviathan-16B-final.gguf -p "你的問題"
授權 (License)
本模型採用 Apache-2.0 授權條款,允許商業與非商業使用,惟須遵守條款規範。
- Downloads last month
- 81
Hardware compatibility
Log In to add your hardware
We're not able to determine the quantization variants.
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support