ShellAI LFM2.5 Bash distillation

Base model: LiquidAI/LFM2.5-350M
Teacher model: LiquidAI/LFM2.5-2.6B
Modifications: LoRA response distillation for natural-language-to-Bash generation,
merged into the base checkpoint and optionally post-training quantized to GGUF Q8_0.
The Q8_0 artifact is not Liquid AI's QAD method and is not labeled as QAD.