LordNeel/Hy3-GGUF
Quantization vs tencent/Hy3. LordNeel/Hy3-GGUF is a GGUF quantization (Q6_K/BF16) of the same weights: 403 tensors match exactly and 561 within quantization error.
Tensors changed: Attention 0/320, lm_head 0/1, MLP 0/321, Norms 0/321, Embeddings 0/1