stac

Hardware Requirements

System Requirements

Minimum Requirements

RAM: 16 GB (32 GB recommended)
GPU: NVIDIA GPU with 8GB+ VRAM (RTX 3080/4080/H100)
Storage: 20 GB free space
CPU: Multi-core processor (Intel i7/AMD Ryzen 7+)

Recommended Requirements

RAM: 32 GB
GPU: NVIDIA GPU with 20GB+ VRAM (RTX 4090/H100)
Storage: 50 GB free space (SSD recommended)
CPU: High-end multi-core processor

Model-Specific Requirements

DistilGPT-2

Conversion Time: 2-5 minutes
VRAM: 4-8 GB
Model Size: ~500 MB

SmolLM2-1.7B-Instruct

Conversion Time: 1-3 hours
VRAM: 20 GB
Model Size: ~3.5 GB

Supported Hardware

Current Support

✅ Software Simulation: Full support on CPU/GPU
✅ NVIDIA GPUs: CUDA 11.8+ or 12.1+
✅ PyTorch: 2.0.0 - 2.5.x

Planned Support (Future Work)

⏳ Intel Loihi-2: Neuromorphic hardware deployment
⏳ BrainChip Akida: Edge neuromorphic processing
⏳ SpiNNaker: Large-scale spiking neural network platform

Installation Notes

CUDA Installation

# For CUDA 11.8
pip install torch==2.3.0+cu118 -f https://download.pytorch.org/whl/torch_stable.html

# For CUDA 12.1  
pip install torch==2.3.0+cu121 -f https://download.pytorch.org/whl/torch_stable.html

# For CPU only
pip install torch==2.3.0+cpu -f https://download.pytorch.org/whl/torch_stable.html

SpikingJelly Installation

# Latest pre-release version required
pip install spikingjelly[cuda] -U --pre

Performance Expectations

Conversion Performance

Simplified Mode: 2-15 minutes
Full Pipeline: 1-3 hours (with quantization and calibration)

Memory Usage

Peak VRAM: 20GB (SmolLM2-1.7B-Instruct)
System RAM: 16-32GB during conversion

Troubleshooting

Common Issues

CUDA Out of Memory: Reduce batch size or use CPU fallback
SpikingJelly Version: Ensure version >= 0.0.0.0.14
PyTorch Compatibility: Use PyTorch 2.0.0 - 2.5.x range

Performance Optimization

Use SSD storage for faster I/O
Close unnecessary applications during conversion
Use simplified mode for initial testing

This site is open source. Improve this page.