dahara1/gemma-4-E4B-it-UD-japanese-imatrix
Quantization (fewer layers) vs unsloth/gemma-4-E4B-it. dahara1/gemma-4-E4B-it-UD-japanese-imatrix matches part of unsloth/gemma-4-E4B-it (677 of its 2130 tensors; the rest of unsloth/gemma-4-E4B-it has no counterpart). On the shared part: dahara1/gemma-4-E4B-it-UD-japanese-imatrix is a GGUF quantization (Q8_0) of the same weights: 381 tensors match exactly and 296 within quantization error.
Tensors changed: Embeddings 0/2, Norms 0/296, MLP 0/126, other 0/85, Attention 0/168