Baekpica/dots3-note-prev-Mixed-Quant-GGUF
Quantization vs dots-studio/dots3-note-prev. Baekpica/dots3-note-prev-Mixed-Quant-GGUF is a GGUF quantization (IQ2_XXS/Q2_K) of the same weights: 140 tensors match exactly and 472 within quantization error.
Tensors changed: Attention 0/235, MLP 0/186, Norms 0/189, lm_head 0/1, Embeddings 0/1