second-state

/

All-MiniLM-L6-v2-Embedding-GGUF

like 17

Follow

Second State 251

I think there's boilerplate text when converting models leads to describing the F16 non-quantized models as "very large not recommend" but this model is a small SBERT model and even the F16 models is only ~50MB. Is there really a need to do quantization on models this small?