stamsam/Instella-Prometheus-GGUF
Quantization vs amd/Instella-MoE-16B-A3B-SFT. stamsam/Instella-Prometheus-GGUF is a GGUF quantization (Q8_0) of the same weights: 108 tensors match exactly and 137 within quantization error.
Tensors changed: Attention 0/54, MLP 0/107, Embeddings 0/1, Norms 0/82, lm_head 0/1