Instructions to use Sculptor-AI/Ursa_Minor_Smashed with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- llama.cpp
How to use Sculptor-AI/Ursa_Minor_Smashed 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 Sculptor-AI/Ursa_Minor_Smashed:F32 # Run inference directly in the terminal: llama cli -hf Sculptor-AI/Ursa_Minor_Smashed:F32
Install from WinGet (Windows)
winget install llama.cpp # Start a local OpenAI-compatible server with a web UI: llama serve -hf Sculptor-AI/Ursa_Minor_Smashed:F32 # Run inference directly in the terminal: llama cli -hf Sculptor-AI/Ursa_Minor_Smashed:F32
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 Sculptor-AI/Ursa_Minor_Smashed:F32 # Run inference directly in the terminal: ./llama-cli -hf Sculptor-AI/Ursa_Minor_Smashed:F32
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 Sculptor-AI/Ursa_Minor_Smashed:F32 # Run inference directly in the terminal: ./build/bin/llama-cli -hf Sculptor-AI/Ursa_Minor_Smashed:F32
Use Docker
docker model run hf.co/Sculptor-AI/Ursa_Minor_Smashed:F32
- LM Studio
- Jan
- Ollama
How to use Sculptor-AI/Ursa_Minor_Smashed with Ollama:
ollama run hf.co/Sculptor-AI/Ursa_Minor_Smashed:F32
- Unsloth Desktop
- Docker Model Runner
How to use Sculptor-AI/Ursa_Minor_Smashed with Docker Model Runner:
docker model run hf.co/Sculptor-AI/Ursa_Minor_Smashed:F32
- Lemonade
How to use Sculptor-AI/Ursa_Minor_Smashed with Lemonade:
Pull the model
# Download Lemonade from https://lemonade-server.ai/ lemonade pull Sculptor-AI/Ursa_Minor_Smashed:F32
Run and chat with the model
lemonade run user.Ursa_Minor_Smashed-F32
List all available models
lemonade list
- Atomic Chat
Download setup-cuda.bat from Sculptor-AI/Ursa_Minor_Smashed: direct link, hf CLI and curl.
- Browser
- Download file 3.38 kB
-
https://huggingface.co/Sculptor-AI/Ursa_Minor_Smashed/resolve/main/setup-cuda.bat
- Command line
-
hf download hf://Sculptor-AI/Ursa_Minor_Smashed/setup-cuda.bat
-
curl -L -o setup-cuda.bat https://huggingface.co/Sculptor-AI/Ursa_Minor_Smashed/resolve/main/setup-cuda.bat
3.38 kB
| @echo off | |
| REM CUDA Setup script for Ursa Minor Smashed model (Windows) | |
| echo π₯ Setting up CUDA environment for Ursa Minor Smashed model... | |
| REM Check if NVIDIA GPU is available | |
| nvidia-smi >nul 2>&1 | |
| if errorlevel 1 ( | |
| echo β ERROR: nvidia-smi not found. Make sure NVIDIA drivers are installed. | |
| pause | |
| exit /b 1 | |
| ) | |
| echo π Checking GPU information... | |
| nvidia-smi | |
| REM Get CUDA version (simplified for Windows) | |
| echo π Please ensure you have CUDA 11.8 or 12.1 installed | |
| REM Create virtual environment | |
| echo π¦ Creating virtual environment... | |
| python -m venv venv-cuda | |
| REM Activate virtual environment | |
| echo β Activating virtual environment... | |
| call venv-cuda\Scripts\activate | |
| REM Upgrade pip | |
| python -m pip install --upgrade pip | |
| REM Install CUDA requirements | |
| echo π Installing CUDA requirements... | |
| echo This may take a few minutes as PyTorch CUDA packages are large... | |
| REM Ask user for CUDA version | |
| echo. | |
| echo Please select your CUDA version: | |
| echo 1. CUDA 11.8 | |
| echo 2. CUDA 12.1 | |
| echo 3. Auto-detect (default) | |
| set /p cuda_choice=Enter choice (1-3, default 3): | |
| if "%cuda_choice%"=="1" ( | |
| echo Installing PyTorch for CUDA 11.8... | |
| pip install torch torchaudio torchvision --extra-index-url https://download.pytorch.org/whl/cu118 | |
| ) else if "%cuda_choice%"=="2" ( | |
| echo Installing PyTorch for CUDA 12.1... | |
| pip install torch torchaudio torchvision --extra-index-url https://download.pytorch.org/whl/cu121 | |
| ) else ( | |
| echo Installing PyTorch for CUDA 11.8 (default)... | |
| pip install torch torchaudio torchvision --extra-index-url https://download.pytorch.org/whl/cu118 | |
| ) | |
| REM Install remaining requirements | |
| echo π Installing remaining dependencies... | |
| pip install numpy>=1.24.0 tiktoken>=0.5.0 tqdm>=4.65.0 | |
| pip install gguf>=0.6.0 sentencepiece>=0.1.99 safetensors>=0.4.0 psutil>=5.8.0 | |
| pip install pynvml>=11.4.1 nvidia-ml-py3>=7.352.0 | |
| pip install matplotlib>=3.7.0 jupyter>=1.0.0 | |
| REM Test CUDA availability | |
| echo π§ͺ Testing CUDA setup... | |
| python -c "import torch; print(f'PyTorch version: {torch.__version__}'); print(f'CUDA available: {torch.cuda.is_available()}'); import sys; sys.exit(0 if torch.cuda.is_available() else 1)" | |
| if errorlevel 1 ( | |
| echo β CUDA not available in PyTorch installation | |
| pause | |
| exit /b 1 | |
| ) | |
| python -c "import torch; print(f'CUDA device count: {torch.cuda.device_count()}'); print(f'Current CUDA device: {torch.cuda.current_device()}'); print(f'CUDA device name: {torch.cuda.get_device_name()}'); print(f'CUDA version: {torch.version.cuda}')" | |
| REM Verify model file exists | |
| if exist "model_optimized.pt" ( | |
| echo β Model file found: model_optimized.pt | |
| ) else ( | |
| echo β οΈ Warning: model_optimized.pt not found in current directory | |
| echo Make sure you have the model file in the same directory as this script | |
| ) | |
| echo. | |
| echo π CUDA setup complete! | |
| echo. | |
| echo π Usage Instructions: | |
| echo To activate CUDA environment: | |
| echo venv-cuda\Scripts\activate | |
| echo. | |
| echo To run CUDA inference: | |
| echo python inference_cuda.py --prompt "Your prompt here" | |
| echo. | |
| echo To run CUDA chat: | |
| echo python chat_cuda.py | |
| echo. | |
| echo To run CUDA benchmark: | |
| echo python benchmark_cuda.py | |
| echo. | |
| echo π Test your setup: | |
| echo python -c "import torch; print('CUDA available:', torch.cuda.is_available())" | |
| pause |