Circuit-Aware Rushd-8B
ุฑูุดูุฏ โ "guidance" or "right path" in Arabic
LoRA adapter for Qwen3-8B, trained with Circuit-Aware methodology to enhance multi-expert reasoning for cybersecurity applications.
Training Details
- Base Model: Qwen/Qwen3-8B
- Rank: 8 (scale: 20)
- Iterations: 500
- Batch Size: 4
- Learning Rate: 1e-5
- Val Loss: 2.148
- Train Loss: 0.030
Capabilities
Enhances 5 Rushd specialists: Router, Logic, Plan, Decision, Game
Hardware compatibility
Log In to add your hardware
Quantized
Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐ Ask for provider support
# Download the model from the Hub pip install huggingface_hub[hf_xet] huggingface-cli download --local-dir Circuit-Aware-Rushd-8B hayulalab/Circuit-Aware-Rushd-8B