mradermacher/nemolita-21b-GGUF
Quantization vs win10/nemolita-21b. mradermacher/nemolita-21b-GGUF is a GGUF quantization (Q8_0) of the same weights: 141 tensors match exactly and 492 within quantization error.
Tensors changed: lm_head 0/1, Embeddings 0/1, Norms 0/141, MLP 0/210, Attention 0/280