baerquants/GLM-5.3-UD-Q4_K_XL-layers
Quantization (fewer layers) vs unsloth/GLM-5.3-GGUF. baerquants/GLM-5.3-UD-Q4_K_XL-layers matches part of unsloth/GLM-5.3-GGUF (18 of its 1809 tensors; the rest of unsloth/GLM-5.3-GGUF has no counterpart). On the shared part: baerquants/GLM-5.3-UD-Q4_K_XL-layers is a GGUF quantization (Q8_0/F32) of the same weights: 7 tensors match exactly and 11 within quantization error.
Tensors changed: Norms 0/6, MLP 0/3, other 0/1, Attention 0/8