qualcomm-ai-hub-community/Gemma-4-E4B-it-Hi-Fi-fraQtl
Quantization (fewer layers) vs google/gemma-4-E4B-it. qualcomm-ai-hub-community/Gemma-4-E4B-it-Hi-Fi-fraQtl matches part of google/gemma-4-E4B-it (677 of its 2130 tensors; the rest of google/gemma-4-E4B-it has no counterpart). On the shared part: qualcomm-ai-hub-community/Gemma-4-E4B-it-Hi-Fi-fraQtl is a GGUF quantization (Q4_K_M) of the same weights: 297 tensors match exactly and 380 within quantization error.
Tensors changed: Embeddings 0/2, Norms 0/296, MLP 0/126, other 0/85, Attention 0/168