tomalamb/Llama-Guard-3-1B-Llama-Guard-3-1B-lora-r1-beavertails-quantized-w4a16

Quantized Rank-1 LoRA Adapter (RTN + Folded)

NOTE: This adapter uses 'Folded' weights for loraB to achieve minimum file size. You **MUST** use a custom loader script to de-quantize and reshape loraB during inference.