mradermacher/bagel-20b-v04-GGUF
Quantization vs jondurbin/bagel-20b-v04. mradermacher/bagel-20b-v04-GGUF is a GGUF quantization (Q8_0) of the same weights: 98 tensors match exactly and 193 within quantization error.
Tensors changed: Attention 0/48, Norms 0/97, MLP 0/144, Embeddings 0/1, lm_head 0/1