studiomaveric/Qwen3.8-27B-UD-Q8_K_XL-layers
Quantization (fewer layers) vs unsloth/Qwen3.8-27B-GGUF. studiomaveric/Qwen3.8-27B-UD-Q8_K_XL-layers matches part of unsloth/Qwen3.8-27B-GGUF (14 of its 866 tensors; the rest of unsloth/Qwen3.8-27B-GGUF has no counterpart). On the shared part: studiomaveric/Qwen3.8-27B-UD-Q8_K_XL-layers is a GGUF quantization (Q8_0/F32) of the same weights: 6 tensors match exactly and 8 within quantization error.
Tensors changed: Norms 0/3, other 0/6, Attention 0/2, MLP 0/3