Architecture choice
#3
by
FlameF0X
- opened
Interesting architecture choice, this model cannot be used as LLM, use it via --model-vocoder in TTS examples
The architecture is not yet merged into upstream llama.cpp (but available on HF), please use runners from https://huggingface.co/LiquidAI/LFM2-Audio-1.5B-GGUF/tree/main/runners