mradermacher/gemma4-e4b-obliterus-custom-GGUF
Same weights, converted (fewer layers) vs Bender1011001/gemma4-e4b-obliterus-custom. mradermacher/gemma4-e4b-obliterus-custom-GGUF matches part of Bender1011001/gemma4-e4b-obliterus-custom (677 of its 2130 tensors; the rest of Bender1011001/gemma4-e4b-obliterus-custom has no counterpart). On the shared part: Same weights stored in another dtype (BF16 → F16/BF16), plus 43 new tensors.
Tensors changed: Embeddings 0/2, Norms 0/296, MLP 0/126, other 0/85, Attention 0/168