lmstudio-community/gemma-3-4b-it-GGUF
Quantization (fewer layers) vs google/gemma-3-4b-it. lmstudio-community/gemma-3-4b-it-GGUF matches part of google/gemma-3-4b-it (443 of its 883 tensors; the rest of google/gemma-3-4b-it has no counterpart). On the shared part: lmstudio-community/gemma-3-4b-it-GGUF is a GGUF quantization (Q8_0) of the same weights: 205 tensors match exactly and 238 within quantization error.
Tensors changed: Norms 0/205, MLP 0/102, Attention 0/136