A 3 billion-parameter Llama 3.2 model fine-tuned end-to-end for general engineering and computer-architecture tasks.
Key Specifications:
Model Size: 3 B parameters
Quantization: 4-bit (INT4)
Frameworks: PyTorch & Hugging Face Transformers
Deployment: On-device inference & cloud
Training data: Engineering textbooks, lecture notes, Logisim circuit diagrams
Why It Matters
Versatile & Accurate
Strikes the sweet spot between footprint and capability—ideal for prototyping, code reviews, and architecture Q&A.
Domain-Focused
Trained on real-world computer-engineering resources (schematics, datasheets, lab write-ups) to deliver actionable, context-aware responses.
Easy Integration
Optimized for edge devices and microservices alike—drop it into your CI/CD pipeline, mobile tooling, or web demo.
For non-commercial use under Meta’s license (see GitHub)
Quick start example
Get Started
đź”— GitHub Repository
For more details on how to use it and for downloading it you can find the model on GitHub which you can find below